AN ABSTRACT OF THE DISSERTATION OF
Christopher R. Friesen for the degree of Doctor of Philosophy in Zoology presented on
December 4, 2012.
Title: Patterns and Mechanisms: Postcopulatory Sexual Selection and Sexual Conflict in a
Novel Mating System
Abstract approved:
Robert T. Mason
Postcopulatory sexual selection—sperm competition and cryptic female choice—has
become a major area of research over the past 40 years. Within this field there are
many outstanding questions at every level of analysis, from proximate to ultimate. The
fitness consequences for both sexes in the period after copulation and before
fertilization are considerable, but are obscured within the female reproductive tract.
Our understanding of postcopulatory mechanisms is especially sparse in taxa other than
birds and insects. Nearly nothing is known in reptiles except that multiple paternity is
common and widespread, and often results from long-term sperm storage across
breeding seasons. We present some of the very first data on the determinants of
fertilization success in the context of sperm competition in reptiles, a group that
accounts for 30% of terrestrial vertebrates.
In the first chapter, “Asymmetric gametic isolation between two populations of
red-sided garter snakes”, we discuss the use of between-population crosses to reveal
gametic isolation. The effect of population density and operational sex ratios on mating
systems and the speciation process has fueled theoretical debate. We attempted to
address these issues using two populations of red-sided garter snakes (Thamnophis
sirtalis parietalis) from Manitoba, Canada. Our study populations differ markedly in their
density mating aggregations, with a 10-fold difference between them. Using
microsatellite markers for paternity analysis of litters produced from within and
between population crosses. We found that the population with highest aggregation
density, and presumably with the highest level of sexual conflict (i.e., when the
evolutionary interests of the sexes differ) over mating, was also the population that
exhibited homotypic sperm precedence. The less dense population showed a distinct
postcopulatory male-size advantage. We also demonstrated that sperm stored within
the female over hibernation can father 20-30% of offspring in a litter.
In the second chapter, “Sperm competition and mate-order effects in red-sided
garter snakes”, we test whether females use mate-order effects to ensure that a larger
(fitter) male will sire her offspring. Does that second male should have precedence in
sperm competition? We tested for second-male precedence using singly-mated females
that mated with a second male. Average proportion of paternity was shared equally
among the first (P1, i.e., proportion of offspring from a litter fathered by the first male to
mate) and second males (P2) to mate, and stored sperm (Pss). This may be a case where
last male precedence breaks down with more than two males. All females were spring
virgins (they had not mated that spring, but may have stored sperm from fall matings);
thus sperm stored presumably from fall matings is important in this system. As the
interval between matings increased P1 increased at the expense of Pss. As the second
male to mate’s copulation duration increased, P1 also increased at the expense of P2.
This last result may indicate female influence over sperm transfer during coerced
matings.
Copulatory plugs (CPs) are found in many taxa, but the functional significance is
debated. Male garter snakes produce a gelatinous copulatory plug during mating that
occludes the opening of the female reproductive tract for approximately two days. In
chapter three, “Not just a chastity belt: the role of mating plugs in red-sided garter
snakes revisited”, we experimentally tested the role of the CPs. In snakes, sperm are
produced in the testes and delivered through the ductus deferens, and the copulatory
plug is thought to be produced by the sexual segment of the kidney and conveyed
through the ureter. We manipulated the delivery of the two fluids separately by ligating
the ducts. We confirmed that the CP is not formed in ureter-ligated males and that
sperm leaks out immediately after copulation. The CP is analogous to a spermatophore.
The protein matrix contains most of the sperm which are liberated as the plug dissolves
within the female’s vaginal pouch.
One of the fundamental principles in sperm competition is that increased sperm
numbers increase the odds of winning in competitions for fertilization success and
males will adjust their ejaculate relative to competition and the quality of his mate. In
chapter four, “Sperm depleted males and the unfortunate females who mate with
them”, we detect significant among-male variation in the number of sperm ejaculated,
and that male mate-order reduces sperm numbers. Male sperm numbers drop
significantly from one mating to the next, and this has implications for sperm
competiveness, as Thamnophis sirtalis exhibits a disassociated reproductive tactic, in
that sperm stores are produced outside the breeding season, and thus cannot be
replenished after mating. Interestingly, however, the on average the mobility of the
sperm increased for a male’s second mating. Therefore, increased sperm quality may
compensate for reduced numbers in a competitive context. Further, females increase
their remating rate when mating with males that are unable to deliver sperm.
In chapter five, “Sexual conflict during mating in red-sided garter snakes as
evidenced by genital manipulation”, we revisited the CP in the context of sexual conflict.
Sex-differences in optimal copulation duration can be a source of conflict, as increased
copulation duration may be advantageous for males as it delays female remating. Males
of many species actively guard females to prevent them from remating, and in some
cases males produce copulatory plugs to prevent remating. If precopulatory choice is
limited at the time of her first mating, conflict may be especially onerous to a female.
The size of the plug is influenced by the copulation duration. We experimentally tested
the contribution of male and female control over copulation duration. We ablated the
largest basal spine on the male’s hemipene and found a reduction in copulation
duration and an increase in the variation of plug mass. Further, we anesthetized the
female’s cloaca and found copulation duration increased, which suggests that males
benefit from increased copulation duration while females actively try to reduce
copulation duration. Therefore, sexual conflict is manifest in divergent copulation
duration optima for males and females.
© Copyright by Christopher R. Friesen
December 4, 2012
All Rights Reserved
Patterns and Mechanisms: Postcopulatory Sexual Selection and Sexual Conflict in a
Novel Mating System
by
Christopher R. Friesen
A DISSERTATION
submitted to
Oregon State University
in partial fulfillment of
the requirements for the
degree of
Doctor of Philosophy
Presented defense December 4, 2012
Commencement June 2013
Doctor of Philosophy dissertation of Christopher R. Friesen presented on December 4, 2012.
APPROVED:
Major Professor, representing Zoology
Chair of the Department of Zoology
Dean of the Graduate School
I understand that my dissertation will become part of the permanent collection of
Oregon State University libraries. My signature below authorizes release of my
dissertation to any reader upon request.
Christopher R. Friesen, Author
ACKNOWLEDGEMENTS
Mom and Dad.
CONTRIBUTION OF AUTHORS
Chapter one, Suzanne Estes conducted the mating trials, and designed the experiment
along with Stevan J. Arnold and Robert T. Mason. Christopher R. Friesen conducted all
statistical and paternity analyses and wrote the manuscript. Chapter 2-5, Christopher R.
Friesen designed the experiments, and conducted all methods and statistical analyses.
The co-authors supported these activities by recording data, conducting mating trials,
and molecular assays.
TABLE OF CONTENTS
INTRODUCTION ................................................................................................................................ 1
ASYMMETRIC GAMETIC ISOLATION BETWEEN TWO POPULATIONS OF RED-SIDED GARTER SNAKE
......................................................................................................................................................... 5
1.1
Introduction ..................................................................................................................... 6
1.2
Methods ......................................................................................................................... 13
1.3
Results ............................................................................................................................ 20
1.4
Discussion....................................................................................................................... 28
SPERM COMPETITION AND MATE-ORDER EFFECTS IN RED-SIDED GARTER SNAKES .................... 36
2.1
Introduction ................................................................................................................... 37
2.2
Methods ......................................................................................................................... 41
2.3
Results ............................................................................................................................ 46
2.4
Discussion....................................................................................................................... 50
NOT JUST A CHASTITY BELT: THE FUNCTIONAL SIGNIFICANCE OF MATING PLUGS IN GARTER
SNAKES REVISITED. ........................................................................................................................ 55
3.1
Introduction ................................................................................................................... 56
3.2
Methods ......................................................................................................................... 63
3.3
Results ............................................................................................................................ 71
3.4
Discussion....................................................................................................................... 79
SPERM-DEPLETED MALES AND THE UNFORTUNATE FEMALES WHO MATE WITH THEM ............ 89
4.1
Introduction ................................................................................................................... 91
4.2
Methods ......................................................................................................................... 95
4.3
Results .......................................................................................................................... 104
4.4
Discussion..................................................................................................................... 110
FEMALE CONTROL OVER COPULATION DURATION AS EVIDENCED BY MALE GENITAL
MANIPULATION AND LOCAL ANESTHETIZATION OF THE FEMALE VAGINAL POUCH IN GARTER
SNAKES ......................................................................................................................................... 118
5.1
Introduction ................................................................................................................. 120
5.2
Methods ....................................................................................................................... 125
5.3
Results .......................................................................................................................... 131
5.4
Discussion..................................................................................................................... 139
THESIS CONCLUSION: FUTURE DIRECTIONS FOR RESEARCH ....................................................... 145
BIBLIOGRAPHY ............................................................................................................................. 148
LIST OF FIGURES
Figure
Page
1.1 Map of Manitoba, Canada…………………………………………………………………………….……………….14
1.2 Box plots showing asymmetric gametic isolation………………………………..…………………………25
1.3 Proportion of offspring sired by the focal male as a function of male size for each separate
cross………………………………………………………………………………………………………..………………………..…27
2.1 Minimum number of fathers per litter…………………………………………………………………………..46
2.2 Proportion of paternity attributed to stored sperm (Pss), the first male to mate (P1), and the
last male to mate (P2)…………………………………………………………………………………..………………….….47
2.3 Before and after line plot showing the effect of male mate-number on fertilization
effects……………………………………………………………………………………………………..…………………….…….48
2.4a Individual regressions of proportion of offspring sired; the effect of the 2nd male to mate’s
copulation duration…………………………………………………………………………………..………………………...49
2.4b Individual regressions of proportion of offspring sired; the effect of the interval between
the first and second matings………………………………………………………………………………..……………..49
2.4c A schematic of the timing of matings relative to ovulation in this species…………....…….49
3.1 Depiction of how the plug is situated in the female’s cloaca, the distribution of sperm within
the plug, and the approximate location of the photographed sections………..…………….…67-69
3.2 Two representative photos of slides wiped across the female’s cloaca within 30 seconds of
the termination of copulation……………………………………………………………………………..……………….72
3.3 Proportion plugs produced given the number of times a male mates………………..………….75
3.4 Scatter plot of copulation durations in matings that did or did not produce plugs..………77
4.1 Regressions plots of the relationship between sperm numbers from natural ejaculates, male
size and copulation duration……………………………………………………………………………..………….…..105
4.2 Before and after line plot of the decrease in sperm numbers from the first to second
matings…………………………………………………………………………………………………………………..…….……105
4.3 Intrapopulational variation in ejaculate quality……………………………………………………………107
LIST OF FIGURES (Continued)
Figure
Page
4.4 Effect of male size on mobility……………………………………………………………………………………..107
4.5 Relationship between male size, first matings and second matings……………………….……108
4.6 Testes mass vs. Male size of 23 Snake Island and 34 Inwood males…………………………….117
4.7 Box plot of residual testes mass of 23 Snake Island and 34 Inwood males…………………..117
4.8 Sperm counts from 14 plugs collected from Snake Island males that mated with Inwood
females late spring 2012 and 25 Inwood males spring 2009………………………………………………117
5.1 The effect of male size on copulation duration and plug mass…………………………………….133
5.2 The effect of copulation duration on plug mass and the relationship between male and
female size among copulating pairs…………………………………………………………………………………..133
5.3 Female treatment effects 2011; copulation duration and plug mass……………………..……135
5.4 Female and male treatment effects 2012 on plug mass……………………………………………...135
5.5 Interaction plot of the combined effects of male and female treatment on plug mass.136
5.6 Interaction plot of the combined effects of male and female treatment on cop. dur….137
5.7 Numbers of sperm present in the copulatory plug compared with the oviduct………..…139
5.8 Interaction plots of oviductal sperm as a function of male and female treatments…….139
LIST OF TABLES
Table
Page
1.1 Population differences based on those females that gave birth and their sires and thus were
used in the paternity analysis…………………………………………………………………………………..............23
1.2 Proportion of offspring sired: within populations vs. between population crosses to assess
the degree of gametic isolation…………………………………………………………………………..…………….…25
1.3 Analysis of the differences between the slopes………………………………………………….…………27
5.1 Model parameters as selected by AIC best model to explain copulatory plug mass...…136
5.2 Model parameters as selected by AIC best model to explain copulation duration….…..137
5.3 Model parameters of oviductal sperm counts…………………………………………….…………….…139
1
PATTERNS AND MECHANISMS: POSTCOPULATORY SEXUAL
SELECTION AND SEXUAL CONFLICT IN A NOVEL MATING
SYSTEM
INTRODUCTION
Theoretical and empirical insights have demonstrated that sexual selection can
generate rapid evolutionary change (Andersson and Simmons 2006). Darwin’s (1871)
concept of sexual selection emphasized an overt “struggle between males” as is seen in
male-male combat or ostentatious ornaments and displays between adults. Since Geoff
Parker’s ground breaking paper (1970), we now recognize that female sexual
promiscuity is common and widespread across animal taxa (Smith 1984; Birkhead &
Møller 1998; Simmons 2001). Therefore, the battle for reproductive success does not
always end with intrasexual competition for mates or intersexual selection by mates
(Darwin 1871; Andersson 1994); instead, sexual selection continues within the
reproductive tract of promiscuous females, where the ejaculates of competing males vie
for fertilization (Parker 1970; Olsson & Madsen 1998) and where females may bias the
contest for fertilization success (Eberhard 1996). Accumulated evidence confirms the
ubiquity and strength of both these forms of postcopulatory sexual selection: sperm
competition and cryptic female choice (Birkhead and Pizzari 2002; Andersson and
Simmons 2006). Yet, little is known about the proximate mechanisms that affect a
male’s sperm quality or patterns of paternity within the female reproductive tract in
taxa other than insects and birds (Wigby and Chapman 2004; Andersson & Simmons
2006; Birkhead et al. 2008). Postcopulatory selection can profoundly affect the
2
evolution of mating systems and species (Parker 1984; Birkhead and Brillard2007). For
example, sperm competition can increase the variance in fertilization success if males
differ markedly in sperm competitiveness (Birkhead and Møller 1998; Shuster and Wade
2003). On the other hand, female multiple mating may also decrease variance in
fertilization success between males by diluting fecundity gains within litters or clutches
(Lorch 2002). This then, may prevent males from reaching their maximum rate of gain.
Reduced variance reduces the strength of sexual selection which may lead to neutral or
even reversed sex roles (Lorch 2002; Shuster and Wade 2003). Cryptic female choice
may also increase variance in male fertilization success and increase strength of sexual
selection by reinforcing precopulatory selection (Eberhard 1996; Simmons 2005).
Through cryptic mechanisms, females may bias paternity toward or away from
particular male phenotypes (e.g., Pilastro et al. 2004). Through its effects on mating
systems, postcopulatory selection can affect speciation rates by affecting sexual
isolation (Coyne and Orr 2004). For example, polyandrous insect clades exhibit higher
species richness and higher incidents of sexual conflict (Arnqvist et al. 2000), which
suggests that female promiscuity may, in some way, lead to sexual conflict (Arnqvist and
Nilsson 2000).
Sexual conflict theory predicts that adaptions and counter-adaptations by the sexes will
arise through a dynamic coevolutionary arms race when one sex limits the evolutionary
interests of the other sex (Parker 1979, 1984; Arnqvist and Rowe 2005). Females are not
evolutionarily passive and have evolved responses to sexual conflict ranging from
3
behavioral adaptations of cryptic female choice such as sperm ejection in fowl (Pizzari
and Birkhead 2000) to morphological adaptations such as the multiple blind-ended
invaginations within the female reproductive tracts of waterfowl (e.g., Brennan et al.
2007; Brennan et al. 2010). Thus, postcopulatory sexual selection and sexual conflict can
be tightly intertwined (Stockley 1997).
Snakes may be an especially fruitful group in which to study postcopulatory processes
and sexual conflict for several reasons. First, they have limited precopulatory choice
and intense male-male competition for females that can lead to sexual conflict. Also, the
incidence of multiple paternity is high, which confirms that most females are sexually
promiscuous. Finally, females can store sperm for long periods which increases the
opportunity for females to bias paternity (Uller et al. 2010). However, for reptiles in
general, and snakes in particular, we know next to nothing about fundamental aspects
of postcopulatory selection such as mate order effects, and the importance of female
sperm storage (Uller and Olsson 2008). Work must be done on these fundamental topics
to lay the ground-work to establish snakes as an important model organism.
Common garter snakes (i.e., Thamnophis sirtalis) have become a model system for the
study of all aspects of reproductive biology (e.g., Blanchard and Blanchard 1941; Fox
1956; Aleksiuk and Gregory 1974; Crews 1984; Mason and Crews 1985; Mason 1992;
Shine 2003; Uller and Olsson 2008). In fact, one of the earliest accounts of multiple
paternity, in any animal, was in garter snakes (Blanchard and Blanchard 1941). The
4
subspecies Thamnophis sirtalis parietalis is remarkable for the large mating aggregations
in the Interlake region of Manitoba, Canada. It is surprising then, that so little is known
of postcopulatory sexual selection in T.s. parietalis. As a consequence of the paucity of
empirical data, descriptive work is a crucial first step to lay a foundation from which
more elaborate and sophisticated studies can be conducted. I hope that this thesis
builds such a foundation. Further, in the conclusions of each chapter and of this thesis, I
suggest several natural extensions of the experimental framework I build. If my modest
suggestions yield fruitful discoveries, T.s. parietalis may become a well-known and
appreciated model system in which to study fundamental questions in postcopulatory
sexual selection and its relation to mating system evolution. For further description of
the mating system of red-sided garter snakes, I direct the reader to Chapter 1.
5
ASYMMETRIC GAMETIC ISOLATION BETWEEN TWO POPULATIONS
OF RED-SIDED GARTER SNAKE
Christopher R. Friesen, Robert T. Mason, Stevan J. Arnold, and Suzanne Estes
Abstract
The role of diverging mating systems in speciation is an area of rich theoretical debate.
Important driving factors of mating system evolution include operational sex ratio and
population density. We address the role of these two factors in a system with the
potential for strong postcopulatory selection and thus gametic isolation: two allopatric
populations of red-sided Garter Snakes (Thamnophis sirtalis parietalis). The most salient
of these differences is in the density of their mating aggregations. We performed
population crosses to assess whether these populations were of reproductively isolated.
The population with highest aggregation density and thus presumably with the highest
level of sexual conflict over mating was also the population that exhibited stronger
homotypic sperm precedence. Further, we found a distinct postcopulatory size
advantage in one population but not in the other. The size advantage was only manifest
in the homotypic cross suggesting intersexual adaptation within that population.
6
1.1
Introduction
Gametic isolation consists of all the reproductive barriers that occur from copulation to
fertilization (Howard 1999; Snook et al. 2009). Successful fertilization relies on complex
interactions between male gametes and the female gametes and reproductive tract,
thus even slight divergence between populations can interfere with fertilization success
(Dziuk 1996; Howard 1999; e.g., Hosken et al. 2002). Moreover, theory predicts that
reproductive traits can rapidly diverge between populations via intersexual
coevolutionary arms races due to sexual conflict (Parker and Partridge 1998; Gavrilets &
Hayashi 2005) and sperm competition (Birkhead and Billiard 2007; Martin-Coello et al.
2009). Consequently, populations that evolve under disparate intensities of sexual
selection and conflict may exhibit some degree of reproductive isolation due to gametic
isolation (Coyne & Orr 2004; Birkhead & Billiard 2007; Martin & Willis 2007; Pitnick,
Wolfner & Suarez 2009). This view is supported by a few empirical studies (e.g.,
reproductive proteins: Swanson & Vacquier 2002; Wolfner 2009; gametic isolation:
Arnqvist 2000; Larson et al. 2012). However, due to its subtle nature during incipient
stages, gametic isolation is often only revealed by asymmetrical patterns of sperm
precedence when sperm competition is induced by researchers (Dziuk 1996; Howard
1999; Marshall et al. 2002; Birkhead & Billiard 2007). Thus, the study of populations that
experience different degrees of sexual selection and/or conflict may reveal how gametic
isolation originates and/or is reinforced and its role in incipient speciation (Marshall et
al. 2002; Birkhead & Billiard 2007). The signature of gametic isolation is clearly manifest
7
in conspecific sperm precedence, which has been extensively documented (Howard
1999; e.g., Wade et al. 1994; Geyer & Palumbi 2005; Larson et al. 2012), but these
barriers are most parsimoniously explained as a secondary result due to ecological
divergence in allopatry (Sobel et al. 2010). Interpopulational sperm precedence due to
gametic isolation is less well studied and may presage incipient speciation, (Robinson et
al. 1994; Howard et al 1998; Marshall et al. 2002; Dixon & Coyne 2003; Eddy 2006;
Birkhead &Billiard 2007). One of the major challenges of this field is to find candidate
wild populations for the study of gametic isolation.
Differences in mating aggregation density and operational sex ratio (OSR) can affect the
opportunity for sexual selection (Emlen & Oring 1977; Jones et al. 2004; Klug et al. 2010)
and the prevalence of sexual conflict (Clutton-Brock & Parker 1995). For example,
females in dense mating aggregations may be harassed into mating with suboptimal
males (Martin & Hosken 2003; Arnqvist & Rowe 2005; Head & Brooks 2006). These
females may then remate to incite sperm competition (Thornhill 1983; Simmons 1987),
mitigate restricted precopulatory choice by bet hedging (Yasui 1997 & 2001; Zeh & Zeh
2001; Fox & Rauter 2003; Simmons 2005), or to gain genetic benefits through their sons
having good/sexy sperm (e.g., Hosken et al 2003). For example, under conditions with
few suitable mates, female sand lizards mate multiply and bias paternity (e.g., sperm
selection, Madsen et al. 1992; Olsson et al. 1996). In theory, sperm selection is a likely
response to limited precopulatory choice due to harassment or coercion as well
(Eberhard 1996; Gowaty 1997; Jones 2002), a view that has received empirical support
8
(e.g., Rivera & Andrés 2002). In populations with smaller or less dense mating
aggregations, the selective regime of conflict and sperm competition may be relaxed
(Arnold & Duvall 1992; Shuster & Wade 2003; Kluge et al. 2010). Thus, populations with
differences in mating aggregation size and density may be good candidates to study
gametic isolation (Wong 2011).
The few studies that address the question of divergence in reproductive success due to
interpopulational differences in mating aggregation size have largely be carried out on
insects such as Drosophila, Tribolium or Sepsis using experimental lines. For example,
Martin and Hosken (2003) experimentally manipulated the degree of sexual conflict in
replicate selection lines of yellow dung flies (Sepsis cynipsea) by adjusting population
density. The high conflict lines diverged most rapidly and showed the highest degree of
reproductive isolation (Martin & Hosken 2003). One of the few studies of gametic
isolation due to sexual conflict in a natural vertebrate population found geographic
variation in sperm production between populations of Trinidadian guppies (Poecilia
reticulata) which lead to interpopulational sperm precedence (Evans & Magurran 2001).
These populations of guppies differed in natural predation rates which in turn
influenced rates of sexual conflict, i.e., forced matings (Elgee et al. 2010).
Red-sided garter snakes of Manitoba, Canada, make excellent subjects for the study of
gametic isolation for several reasons. First, garter snakes in general exhibit multiple
paternity and long term sperm storage (Uller and Olsson 2008), so postcopulatory
9
mechanisms should be of evolutionary importance. Second, relatively large sample sizes
can be readily obtained at den sites during the spring breeding season. Third, males
display robust courtship in controlled arena trials so the last male to mate is easily
identified for sperm precedence studies. Fourth, once they have mated females are
easily transported back to laboratory facilities to give birth in late summer, so that
offspring can be collected and maternity is certain. Finally, mating aggregation densities
vary among populations.
Although multiple paternity is well documented in garter snakes (e.g., Blanchard &
Blanchard 1943; Wusterbarth et al. 2010; reviewed in Olsson & Uller 2008), the identity
of the last potential father is unknown in these studies. Without knowing the identity of
the last potential sire, neither order effects on sperm precedence nor the prevalence of
stored sperm usage can be established. Again, knowing the identity of the last potential
male is of particular importance when considering postcopulatory selection in snakes as
there are many examples of long term sperm storage throughout the group (Olsson &
Uller 2008; Uller et al. 2010). It is crucial to first establish patterns of paternity and
sperm precedence if we are to then work on the more difficult task of untangling
mechanisms by which precedence occurs, which is the contemporary focus in other,
better studied taxa (Pitnick et al. 2009; Wong 2011). Here we present data from the first
study in snakes that uses experimental crosses to assess gametic isolation in two
populations with very different aggregation densities.
10
Red-sided garter snakes mating system
In central Manitoba, Canada, large numbers of red-sided garter snakes overwinter in
limestone sinkholes from which they emerge in late April (Aleksuik & Stewart 1972) and
to which many faithfully return each fall (Gregory 1974; Macmillan 1995). Manitoba T.
sirtalis display a scramble competition polygyny mating system (Thornhill & Alcock
1983). Males surface first, and form very large and dense aggregations around the
hibernaculum (den) for 4-5 weeks (Gregory 1975; Shine et al. 2001). In contrast, females
surface and remain around the den for an average of 3-4 days before they migrate to
summer feeding grounds (Shine et al. 2001). The intersexual difference in spatial and
temporal distribution of the two sexes generates extremely male-biased OSRs (Emlen &
Oring 1977). Males are not aggressive toward one another, but instead compete to find
emerging females where they form “mating balls” – aggregations of males courting and
vying for copulation with a single female (Gregory 1974).
Mating aggregation size has the potential to affect sexual selection in this species. For
example, differences in aggregation density translate into differential male mating
success based on male size. Field data collected from a den with very large mating
aggregations (3 – 62 males; Shine et al. 2006) reveals that in some years mating is
completely random (2 of 5 years) with respect to male size (Shine et al. 2000c). Even in
those years when larger males had a significant advantage, the difference in mean
snout-to-vent length between successful and unsuccessful males was quite small (< 1
cm; Shine et al. 2000c). However, in experimental mating trials with small aggregations,
11
larger males have a strong mating advantage over smaller males (< 20 males; Shine et al.
2000c). Whether it is a function of male size advantage or female choice, sexual
selection on male body size operates differently at different aggregation densities.
Sexual conflict is potentially a more influential evolutionary factor in the larger den sites
with large mating aggregations (Shine et al 2000a; Shine et al. 2004; Shine et al. 2005;
Shine et al. 2005). Females in the largest mating balls risk exhaustion, harm or death by
crushing under the mass of courting males (Shine et al. 2004). In addition, precopulatory
female choice may be limited within the largest mating balls where male-to-female
ratios are most strongly skewed toward males (20-62 males per female). Females might
be coerced to mate (i.e., cryptic forcible insemination; Shine et al. 2003; Shine et al.
2004). In these circumstances theory predicts that populations experiencing sexual
conflict should experience female-male antagonistic coevolution (Gavrilets 2000;
Gavrilets & Hayashi 2005). Males should have better fertilization success in the context
of sperm competition against rival males from populations that have lesser degrees or
no sexual conflict. Further, male advantage should be most evident when males mate
with coevolved females. Male-female coevolution would produce a measurable degree
of gametic isolation between populations, because runaway evolution of reproductive
traits occurs more rapidly as population size increases (Gavrilets 2000; e.g., Martin &
Hosken 2003). Such gametic isolation would be revealed by paternity analysis of litters
produced by crosses between populations with difference mating densities.
12
13
1.2
Methods
In this study, we tested for patterns of gametic isolation between two populations with
different mating densities. We predicted that stored sperm from males from the
population with lower density mating aggregations would do better when pitted against
the population with lower density mating aggregations. The trend should be most
pronounced within the females from the high density population because of
antagonistic coevolution. However, if sperm competition is extremely intense in the
larger aggregations, then the males with that aggregation density should do relatively
better in both populations, assuming sperm number is the predominate mechanism by
which males improve sperm competitive ability in these populations.
Study Populations
Animals were collected from two sites (Figure 1.1 Map): Inwood animals (hereafter
abbreviated I) collected from a large communal den 1.5 km north of Inwood in central
southern Manitoba (50° 31.58’N 97°29.71’W); Snake Island animals (hereafter
abbreviated S) were collected from small aggregations distributed along a low ridge on
Snake Island in Lake Winipegosis, in central southern Manitoba (51° 38.53’N 99°
49.42’W). The Snake Island snakes hibernate and mate on the island but forage on the
mainland (Mason et al. 1991; CRF, SE, RTM pers. obs.).
14
Figure1.1: Map of Manitoba, Canada.
The stars represent our two study
populations. Snake Island is
approximately 250 kilometers
northwest of the Inwood population.
The reproductive biology of the Inwood population has been studied extensively and
consists of approximately 35,000 snakes reported in a previous season (Shine, Langkilde
et al. 2006), most of which emerge from a single depression approximately 30 M x 10 M.
Females at the Inwood den site emerge into masses of snakes numbering in the
thousands and where mating balls ranging from 20-62 males gather around the females
(Shine, Elphick et al. 2001). Most females mate before leaving the main aggregation in
the base of the den (Whittier & Crews 1986). We do not have exact census data for the
Snake Island population but it is approximately an order of magnitude smaller. Snakes
emerge along a short 1-2 M high compression ridge that runs the length of the island.
15
The Snake Island population is comprised of over 100 but not more than 500 individuals
(Mason et al. 1991). The overall density of this population is < 1% that of the Inwood
population and mating aggregations were rarely observed with more than 15 males and
one female (Mason et al. 1991; Estes, Mason & Friesen personal observations during
spring 2004 & 2012).
Female collection
Females were collected as they reached the surface at the den during the spring of
2004. It is unlikely that females mate over the winter as body temperatures are very low
(Lutterschmidt et al. 2006). Female Thamnophis sirtalis are known to mate during
autumn (Blanchard & Blanchard 1943), and can potentially store sperm for long periods
including the previous spring (Stewart 1972). Long-term sperm storage is important for
our study, as most females will have mated previously and thus carry sperm within their
reproductive tracts prior to our controlled matings. Any paternity we cannot assign to
our focal male is most likely due to stored sperm and thus necessarily from withinpopulation matings (homotypic). Stored sperm (homotypic) should be at a numerical
disadvantage, but if mechanisms of homotypic sperm precedence have evolved, then
paternity may still be biased towards homotypic stored sperm when a female has mated
with a heterotypic male (a male from outside her population).
Mating trials
Mating trials were conducted at the Chatfield Research Station, Manitoba, Canada,
within a few days of capture. A mating trial consisted of placing females in 1Mx1Mx1M
16
nylon semi-natural arenas with 10 males. Thus, the OSR for the mating trails was held at
a moderate and realistic level for all matings (Shine et al. 2001). Four combinations of
population-pairings were used in mating trials: Inwood female mated with Inwood male,
Inwood female mated with Snake Island males, Snake Island females mated with Snake
Island males, and finally Snake Island females mated with Inwood males (hereafter
abbreviated: I x I, I x S, S x S, S x I respectively). Note that in each cross the first
population is that of the female and the second is the male’s population.
Once a female was placed in the arena, she was observed until mating occurred and the
pair was carefully removed from the arena before copulation terminated and then
placed in a smaller area alone so that the focal male was known with certainty. When
copulation was terminated the animals were weighed and measured, and small (< 10
mm) tissue samples were taken from the tail of the male and female (Garner et al.
2004). Tissue was placed directly in Nunc tubes ¾ filled with Drierite desiccant where
they remained until DNA extraction. Females were clipped with unique markings on
their ventral scales to identify them through the next 12 months they would be kept in
captivity (Blanchard 1933). The males were released at the point of capture after tissue
samples were taken.
Husbandry
Females were transported to Oregon State University, housed in 38 L aquariums, given
water ad libitum and fed worms (Lumbricus terrestris) or fish (Oncorhynchus
17
tshawytscha) weekly. The females gave birth in the lab from mid-August through early
September. Seventy-one of the one hundred and ten mated females gave birth (64.5%).
Female garter snakes are capital breeders and most likely breed every other year
(Gregory 2006; RTM per. obs.) so this is close to the expected proportion of births.
Offspring were weighed, measured and tissue was collected in the same manner as the
adults. Differential offspring mortality before birth is one factor that can skew results in
studies of sperm precedence is (Zeh & Zeh 1997; Tregenza & Wedell 2000). To assess
this possibility, we tested for differences in litter size, number of stillbirths and birth
rates between homotypic and heterotypic crosses after controlling for female size which
has a known, direct effect on litter size.
Molecular methods
DNA was extracted from tail tip tissue (Garner, Pearman et al. 2004) in 200 µl 5% chelex
1% ProK solution incubated at 56°C for 2 hours and then 8 minutes at 100°C. We used
three microsatellite loci to exclude the focal male from paternity: Ts1 (McCracken et al.
1999) and Nsµ2 and Nsµ3 (Prosser et al. 1999). All three loci were multiplexed in a
single 12 µl PCR reaction (1 µl template; 6.25 µl Multiplex mix (Qiagen cat no. 206143);
1.25 µl 2 µM of each primer (Invitrogen); 3.5 µl molecular grade H20). PCR reactions
were carried out in a Biorad thermal cycler (C1000). Amplification conditions consisted
of 15 min at 95°C, followed by 35 cycles of 30 sec at 95°C and 1:30 min at 58°C, and 1:30
min at 72°C followed by a 10 min at 72°C elongation phase. Reaction products were run
in an ABI 3100 genetic analyzer, and the alleles were visualized using ABI Genotyper
18
software. Peaks were assigned manually, and each offspring was checked against the
mother’s genotype. Any offspring that did not have a maternal allele was rerun along
with the mother’s genotype and a random subset of siblings to check for errors. Seven
families were eventually excluded from analysis either because of apparent maternal
null alleles (N = 2), trouble amplifying maternal template DNA (N = 4) or an apparent
mismatch between a mother and her litter (N = 1). A total of 66 families (1409 offspring)
were successfully genotyped.
We used an exclusion based protocol for paternity assignment. We tested for HardyWeinberg Equilibrium (HWE) and estimated average exclusion probability using CERVUS
3.0.3. (Marshall et al. 2007). The three loci chosen for this study were all highly
polymorphic and loci Ts1 and Nsµ3 were in HWE (Ts1, χ2 = 298.9, d.f. 325, P = 0.847;
and Nsµ3, χ2 = 50.386, d.f. 66, P = 0.923) however, locus Nsµ2 was not (Nsµ2, χ2 =
503.5, d.f. 190, P < 0.001). The average exclusion probability, based on the genotypes
from 56 random adults, for each locus was: Ts1 (1 – 0.1160) = 0.884 exclusion
probability (second parent), with 26 alleles found in 56 genotyped individuals; H0
=0.9643; Nsµ2 (1 - 0.1654) = 0.8346 exclusion probability (second parent), with 20
alleles found in 56 individuals, H0 = 0.5357; Nsµ3 (1 - 0.2508) = 0.7492 exclusion
probability (second pair), with 12 alleles found in 56 individuals, H0 = 0.8929. The
combined average exclusion probability using all three alleles is (1-0.0048) = 0.9952 and
any pair of loci yields greater than 0.95 confidence of correctly excluding our focal male.
In many cases we relied on only two loci because one of the three loci was
19
uninformative (e.g., mother and focal male shared alleles at that locus). The single locus
exclusion probability of Ts1 calculated using all the adults from this study was 0.98,
provided the maternal and focal male’s genotypes did not match. A conservative
estimate of the minimum number of fathers per litter was calculated by dividing the
number of paternal alleles by 2 after assigning alleles to the known male. We added one
male to this number to account for the known male.
Statistical analyses
All statistical analyses were conducted using either Sigma plot 11.0 (Chi-square, ANOVA
and T-tests), except the ANCOVA that was conducted using XLSTAT and Generalized
linear regression in PASW 17.0. Body Condition Index (BCI) was calculated as the
residual derived from the quadratic regression of female mass on female SVL fitted in
Sigma Plot 11.0 (Adj. R2 = 0.918; Note the Adj. R2 for SLR model = 0.827). Proportion of
offspring sired was arcsine-square root transformed, and male snout-to-vent length
(SVL) was natural log transformed to correct for unequal variance among crosses for
regression analyses. We tested for differences in homotypic sperm precedence using
one-tailed t-tests because we made the a priori prediction that there would be a bias
toward stored sperm usage (homotypic) when females were mated with a heterotypic
male regardless of the source of male-female coevolution.
20
1.3
Results
Population comparisons
Of the animals that gave birth, Snake Island females were significantly longer than
Inwood females (SVL: t-Test t64 = 3.675, P < 0.001), but not significantly heavier;
(Ln(Mass): t-Test t63 = 1.70, P = 0.094). Accordingly, Inwood females had higher Body
Condition Indexes (BCI) than Snake Island females (BCI: t-Test t63 = 3.41, P = 0.001). Of
the potential sires, Snake Island males were significantly longer and heavier than
Inwood males (SVL: t-Test t63 = 6.47, P < 0.001; Mass: t-Test t63 = 3.443, P < 0.001).
However, Inwood males were in better condition than Snake Island males mean (BCI: tTest t63 = 2.594, P = 0.012), summarized in Table 1.1.
Post zygotic isolation
Birth Rate
Larger females are more likely to give birth (Logistic regression: Likelihood Ratio Test
Statistic: 33.203, P < 0.001). Female body condition differed significantly between
females that did or did not give birth and between population (two-way ANOVA; Female
population F1, 106 = 22.01, P < 0.001; Gave Birth F1, 106 = 10.23, P = 0.002; interaction F1, 106
= 0.429, P = 0.514). Inwood females that gave birth had higher BCI than those that did
not, but the same relationship did not exist for Snake Island females (Holm-Sidak
method for multiple comparisons: I, t = 3.03, unadjusted P = 0.003; S t = 1.65,
unadjusted P = 0.103).
21
To test whether females were less likely to give birth when mated with a heterotypic
male we analyzed birth rate among crosses (Table 1.1). The proportion of females that
gave birth did not differ significantly among crosses (Chi Square test; X2 = 3.355 with 3
d.f., P = 0.340) nor between female population (Chi Square test; χ2 = 1.327 with 1 d.f., P
= 0.249). We also compared within and between population crosses directly, but still
found no differences in birth rate between homotypic and heterotypic matings (I x I vs. I
x S Chi Square test; χ2 = 1.195 with 1 d.f., P = 0.274; S x S vs. S x I Chi Square test; χ2 =
0.294 with 1 d.f., P = 0.588).
Litter size and stillbirths per cross
To assess whether genetic incompatibility and embryo reabsorption might have skewed
paternity patterns we tested for differences in litter size and the number of stillbirths
among the crosses. Without controlling for female size, litter size was not different
among crosses (F3, 71 = 0.577; P = 0.632). As noted earlier, females differ in size between
populations and female mass predicts litter size (SLR F1, 71 = 67.168, P < 0.001).
Therefore, we compared residuals of litter size given female mass of between
homotypic and heterotypic crosses in females from same populations and found no
differences (I x I vs. I x S, one-tailed t-test; t34 = 0.005, P = 0.498; S x S vs. S x I one-tailed
t-test; t35 = 0.134, P = 0.447). The number of stillbirths per cross yielded no support for
postzygotic sexual isolation. Across all females mean proportion of stillbirth offspring
was 0.073 (95% C.I. of mean = 0.023). Mean proportion of stillbirth offspring was not
affected by Cross (Kruskal-Wallis one factor ANOVA, Hdf=3 = 1.844, P = 0.605) nor did a
22
comparison between homotypic and heterotypic crosses (I x I vs. I x S Chi Square test; χ2
= 2.057 with 1 d.f., P = 0.151; S x S vs. S x I Chi Square test; χ2 = 0.011 with 1 d.f., P =
0.915). Of those litters we were able to successfully assign paternity, we tested whether
stillbirth offspring were more likely to have been fathered by heterotypic males and
found no differences (I x I vs. I x S Chi Square test; χ2 = 1.637 with 1 d.f., P = 0.201; S x S
vs. S x I Chi Square test; χ2 = 0.2.057 with 1 d.f., P = 0.151). There was no evidence of
differences in litter size or incidence of still births among crosses.
23
Table 1.1: Population differences based on those females that gave birth and fathers used in the
paternity analysis.
N
df
Mean(SEM)
t
p-value
Inwood
32
64
106.9 (5.26)
1.69
0.094†
Snake Is.
34
Inwood
32
3.67
<0.001
Snake Is.
34
Inwood
32
3.68
<0.001
Snake Is.
34
Inwood
35
3.44
0.001
Snake Is.
31
Inwood
35
6.46
<0.001
Snake Is.
31
Inwood
35
2.59
0.012
Snake Is.
31
I♀ x I♂
18
0.64
0.591
I♀ x S♂
14
23.9 (1.39)
SC x I♂
17
20.8 (1.81)
S♀ x S♂
17
20.4 (2.07)
Female Mass (g)
121.6 (6.57)
Female SVL (cm)
64
65.3 (1.15)
71.2 (1.12)
Female BCI
64
5.81 (2.20)
-5.63 (2.2)
Male Mass (g)
64
30.4 (1.25)
38.7 (2.17)
Male SVL (g)
64
45.7 (0.87)
52.2 (0.95)
Male BCI
64
1.05 (0.52)
-1.22 (0.72)
Litter size x Cross
3, 61
21.2 (2.08)
♀ gave birth/mated pairs (%)
Birth rate x Cross
I♀ x I♂
21/39 (53%)
I♀ x S♂
15/22 (68.2%)
S♀ x I♂
18/24 (75%)
S♀ x S♂
17/25 (68%)
† t-stat and p-value generated on ln transformed data
24
Paternity
Minimum number of sires per litter
Eighty-five percent of litters exhibited multiple paternity, and there was no difference
among crosses in the proportion of litters with multiple paternity (χ2df 3 = 3.901, P =
0.272). The median minimum number of fathers per litter was 2.0 among all crosses (1-3
fathers). There is no difference among the crosses in the median number of fathers
within a litter (Kruskal-Wallis one factor ANOVA on ranks, H= 0.823 (d.f. = 3), P = 0.844).
Litters with more fathers were not significantly larger with mean litter sizes as follows: 3
fathers, 21.83 offspring; 2 fathers 22.32; 1 father 16.78 (one factor ANOVA, F 2, 62 =
1.983, P = 0.146). Larger females did not have more fathers per litter (SVL: one factor
ANOVA, F 2, 62 = 0.775, P = 0.465).
Gametic isolation
Mean proportion of offspring fathered by the known male across all crosses was 0.67 ±
0.059, significantly greater than 0.50, which is what we would expect under an equal
weighting of paternity (two-tailed test, t65 = 5.595, P < 0.001). Mean proportion of
offspring fathered by the known male within crosses (± SE) were as follows I x I = 0.754 ±
0.045, I x S = 0.617 ± 0.060, S x I = 0.629 ± 0.052, S x S = 0.643 ± 0.081 (Table 1.2 and
Fig. 1.2). Inwood females gave birth to more offspring attributable to homotypic stored
sperm when they mated with heterotypic males (I x I vs. I x S: one-tailed t-Test, t 30 =
1.866, P = 0.036). The Snake island female crosses (S x S and S x I) did not have equal
variance in proportion of offspring fathered (Levine’s Test F1, 31 = 6.123, P = 0.019).
25
However, there was no difference in the proportion of offspring attributable to stored
sperm between heterotypic and homotypic crosses (S x S vs. S x I: one-tailed MannWhintey rank-sum test H = 0.148, P = 0.442). The increased variance within the S x S
cross is probably due to the effect of male size on the probability of paternity that
occurs solely within this cross (see section below).
P = 0.442
P = 0.036
1.0
Proportion of offspring sired (P2)
Proportion of offspring sired (P2)
1.0
0.8
0.6
0.4
0.2
0.8
0.6
0.4
0.2
0.0
0.0
IxI
SxI
IxS
SxS
Cross (Female x Male)
Cross (Female x Male)
Figure 1.2: Box plots showing asymmetric gametic isolation; more stored sperm from homotypic males
was used when Inwood females mate with Snake Island males. The horizontal dashed lines within the
boxes are means, and the horizontal solid lines are medians. The boxes enclose 50% of the data, and the
whiskers are 1.5 interquartile ranges, the solid line is the median and the dashed line is the mean. There is
more variance within the Snake Island by Snake Island cross.
Table 1.2: Proportion of offspring sired (P): within populations vs. between population crosses to
assess the degree of gametic isolation.
Cross
n
df
P (±SE)
t
p- value
I♀ x I♂
18
30
0.754 (0.045)
1.866
0.036
I♀ x S♂
14
S♀ x I♂
17
0.25
0.442
S♀ x S♂
17
0.617 (0.063)
32
0.643 (0.081)
0.629 (0.052)
26
The effect of male size on paternity
Longer Snake Island males had significantly higher paternity when they mated with
Snake island females, but male size did not have an effect on paternity in any other
cross (Generalized linear regression in PASW 17.0; Male SVL Wald Chi-Square = 1.58
with 1 d.f., P = 0.209; Snk x Snk Cross, Wald Chi-Square = 10.55 with 3 d.f., P = 0.001; Snk
x Snk Cross : Ln (Msvl) interaction, Wald Chi-Square = 10.44 with 1 d.f., P = 0.001).
ANCOVA also corroborated this statistical result (Full model ANOVA F7, 63 = 2.714, P =
0.017; Ln(Msvl)*Cross interaction P = 0.005) and provided a comparison of slopes
(Figure 1.3 and Table 1.3). Female size had no effect on paternity (Female SVL Wald ChiSquare = 0.392 with 1 d.f., P = 0.531; Snk x Snk Cross, Wald Chi-Square = 2.458 with 3
d.f., P = 0.483; interaction Wald Chi-Square = 2.580 with 1 d.f., P = 0.461). Note that
these data were transformed to account correctly for unequal variance among the
crosses. Paternity increases significantly with male size in only the S x S cross.
27
Proportion of offspring fathered by focal males
1.6
A
1.4
Arcsin(Sqrt(P2))
SxS
1.2
B
1.0
B
IxI
IxS
SxI
0.8
0.6
B
0.4
0.2
3.4
3.6
3.8
4.0
4.2
Ln (Msvl)
Figure 1.3: Proportion of offspring sired by the focal male (Arcsine Sqrt. (P2)) as a function of Ln(male SVL)
for each separate cross. The filled-in circles represent I♀ x I♂ cross, the open circles represent the I♀ X S♂
cross, the filled-in diamonds represent the S♀ x S♂ cross, and the open diamonds represent the S♀ x I♂
2
cross. The equations for the lines, Adj. R and the P-value of separate regression analyses for each cross
2
are as follows: Arcsin(sqrt(P2))-IxI = 0.269 + (0.216 * Ln(Msvl(IxI))), Adj. R = 0.000, P = 0.738;
2
Arcsin(sqrt(P2))-SxS = -7.668 + (2.196 * Ln(Msvl (SxS))), Adj. R = 0.262, P =0.025; Arcsin(sqrt(P2))-IxS =
2
0.152 + (0.198 * Ln(Msvl (IxS))), Adj. R = 0.000, P = 0.821; Arcsin(sqrt(P2))-SxI = 3.763 - (0.740 * Ln(Msvl
2
(SxI))), Adj. R = 0.058, P = 0.180. ANCOVA analysis of differences in slopes is presented below in Table
1.3 and the letters to the right of the regression lines indicate which lines are significantly different from
one another.
Table 1.3: Analysis of the differences between the slopes. There is a positive relationship between
the probability of paternity and male size in the S x S cross; Ln(Msvl)*Cross interaction P = 0.005)
Comparison
β1- β2
t
Pr > |t|
Significant
SxS vs IxI
2.303
2.373
0.021
Yes
SxS vs IxS
2.322
2.118
0.039
Yes
IxI vs IxS
0.019
0.017
0.987
No
SxS vs SxI
3.259
3.728
<0.0001
Yes
IxI vs SxI
0.956
1.076
0.287
No
IxS vs SxI
0.938
0.915
0.364
No
28
1.4
Discussion
Reptiles, and snakes in particular, are vastly underrepresented in studies of
postcopulatory sexual selection; what little we know within reptiles is largely limited to
testing for multiple paternity (Uller & Olsson 2008; Uller et al. 2010). This is the first
study in snakes that addresses sperm precedence patterns in a natural population. We
demonstrate last male sperm precedence, and also show that stored sperm from
previous seasons fertilize a substantial proportion of the offspring from any given litter.
Our data also demonstrate asymmetry in gametic isolation between geographically
distinct populations of garter snake with different social contexts for mating; i.e., mating
aggregation densities. The asymmetry cannot be explained by genetic incompatibility
between populations, but must be a consequence of a postcopulatory, prezygotic
mechanism. Further, larger Snake Island males have a distinct advantage in
postcopulatory selection, but only when they were mated with females from their own
population.
Sperm precedence patterns and multiple paternity in Thamnophis
Our data are consistent with five previous studies using microsatellites within the genus
Thamnophis that found 37.5-100% of litters exhibited multiple paternity (McCracken et
al. 1999; King et al. 2001; Garner et al. 2002; Garner, Pearman et al. 2004; Wusterbarth
et al. 2010; reviewed in Uller & Olsson 2008). The occurrence of multiple paternity
clearly indicates the opportunity for postcopulatory selection within these populations.
Furthermore, the frequency of multiple paternity gauges the strength of postcopulatory
29
selection, i.e., the risk that a male’s sperm will compete with the sperm of other males.
Eighty-five percent of the litters showed multiple paternity in our system. This is
probably an underestimate of female remating rates as some males may father all
offspring even in the face of sperm competition due to chance, superior ejaculate
quality and/or cryptic female choice. In these populations, the minimum level of sperm
competition risk is high given that these females were only allowed to mate once in the
spring and thus the main competition came from sperm stored in previous seasons.
Another way to characterize sperm competition is the intensity or the number of
competitors’ ejaculates a male is expected to encounter. In some cases, we found that
three fathers were required to explain paternity and there was an average of two males
per litter, which implies a moderate intensity of sperm competition (Parker 1998).
Mate-order effects
Mate-order effects within the spring mating season and the prevalence of multiple
mating in the spring were not addressed in this study. However, our data provide some
insight into sperm precedence due to mate-order effects. Without considering the effect
of “cross”, there was moderate to high second-male precedence, as the last male to
mate fathered an average of 67% of the offspring, significantly higher than 50%
paternity we might expect if all males had an equal chance at paternity (Simmons 2001).
If this figure accurately estimates order precedence for temporally proximate vernal
multiple matings, it would complicate explanations for the evolution of mating plugs
which only inhibit remating for 48 hours (Devine 1975; Shine et al. 2000b). However,
30
more elaborate experiments are needed to address mate-order effects of vernal
multiple matings, the role of the copulatory plug, and female mechanisms that may bias
paternity.
Sperm storage
Thamnophis have well developed sperm storage crypts (Fox 1956; Hoffman & Wimsatt
1972) and can store sperm for long periods (e.g., Stewart 1972). Female sperm storage
increases the chances that sperm form different males will compete for fertilization of
ova.
Our data demonstrate that sperm stored from previous seasons can be used to fertilize
ova and play a potentially important role in postcopulatory selection (Uller & Olsson
2008; Uller et al. 2010). In the Interlake region (e.g., Inwood), and presumably in the
Snake Island population as well, matings occur in the autumn prior to winter hibernation
as well as the spring (Blanchard & Blanchard 1941 & 1943; Aleksiuk & Gregory 1974;
Whittier & Crews 1986; CRF pers. obs. autumn 2008 & 2009). Because all females
collected for this study were collected immediately upon emergence and had no
opportunity to mate during the spring except in our controlled mating trials, our study
provides strong evidence for the effectiveness of autumnal matings or matings from the
previous spring. In a study of the inheritance of melanism, Blanchard & Blanchard (1941)
found convincing evidence that autumn matings can produce viable offspring the
following summer (over 10 months later). Stewart (1972) attributed offspring born 53
31
months after the last possible opportunity for mating to stored sperm, so it is possible
that matings during the previous spring could account for some of the extra-sire
paternity we found. But we cannot distinguish between sperm used from autumn
matings and matings from the previous spring. Although recently published examples of
facultative parthenogenesis in snakes may call into question stored sperm usage across
seasons (Booth et al. 2012), our study found none of the hallmarks of
parthenogenetically produced litters i.e., unusually small litters composed entirely of
the homogametic sex (males in this case) (Booth et al. 2012).
The high frequency of multiple paternity in our study, coupled with minimal
precopulatory choice and female sperm storage sets the stage for the evolution of
cryptic female choice (CFC) in these snakes (Devine 1984; Uller et al. 2010). More work
needs to be done to identify mechanisms of postcopulatory selection in general and
cryptic female choice specifically. It seems Thamnophis are ideal candidates as models
for such research given that long term sperm storage is well documented. In fact, our
results support the possibility of cryptic female choice.
Gametic Isolation between populations of Thamnophis
Our experimental design took advantage of the fact that stored sperm would
necessarily come from homotypic males. We predicted that we would observe
homotypic sperm precedence if postcopulatory prezygotic barriers had evolved
between conspecific populations (Marshal et al. 2002; Birkhead & Billiard 2007). We
32
found evidence of homotypic sperm precedence, but only within Inwood females. Given
that sperm from autumnal matings would likely be in a disfavored role due to sperm
attrition, our result seems remarkable and suggests that the asymmetry would be more
pronounced if there was parity in sperm numbers. Parity in sperm numbers can only be
assured in artificial insemination trials, but would reduce or eliminate
precopulatory/copulatory cues that females might use to bias paternity via sperm usage
or differential sperm transport (Eberhard 1996). Natural vernal double-matings using
the same crossing design would be useful in further elucidating sperm precedence
patterns we uncovered.
Reduced in paternity in between-populations crosses is not necessarily the result of
gametic isolation, but can result from postzygotic factors such as genetic incompatibility
(Zeh & Zeh 1997; Stockley 1999; Tregenza & Wedell 2000; Birkhead & Billiard 2007).
However, we did not find evidence of postzygotic isolation. The proportion of stillbirths
was not different for heterotypic vs. homotypic matings and the number of stillbirth
offspring attributable to heterotypic males was no different than for homotypic males.
Neither the proportion of females that gave birth nor litter size differed significantly
after controlling for female size among population crosses. Thus, the hypothesis that
offspring were aborted and reabsorbed before birth lacks support. Thus the most likely
explanation for the observed asymmetry in paternity manifests after copulation but
before fertilization, i.e., gametic isolation via homotypic sperm precedence. The fact
that sperm precedence of Inwood males did not translate to Snake Island females
33
suggests female-male coevolutionary processes may account for asymmetry rather than
the Inwood males being generally superior sperm competitors and is suggestive of
intersexual coevolution but could also occur via female preferences for homotypic
sperm.
Red-sided garter snake populations recolonized Manitoba less than 12,000 years ago as
they have only been able to inhabit the area after the last glacial retreat (Rye 2000;
Dyke 2004; Placyk et al 2007). Within the Interlake region, these snakes have a
discontinuous spatial population structure based on den-philopatry (Gregory 1974 &
1977; Macmillan 1995; LeMaster & Mason 2003; Mooi, Wiens & Casper 2011). In
general, asymmetry in reproductive isolation has been modeled as the transitory effect
of rapid divergence of sexually selected traits facilitated by drift along the stable line of
equilibria (Arnold et al. 1996; Uyeda et al. 2009). However, given the lack of evidence
for population differentiation of neutral markers in Thamnophis sirtalis in general (DiLeo
et al. 2011) and these populations of T.s. parietalis sin particular (Westphal 2007 PhD
Dissertation) drift may not fully account for our results. Den-philopatry, dispersal and
environmental differences probably have consequences for the natural history of these
animals and thus the direction and strength of natural selection may differ between
populations. However, there is growing acceptance of the idea that sexual selection can
drive speciation (Coyne & Orr 2004; Ritchie 2007; Maan & Seehausen 2011).
34
Divergence between the two populations due to sexually antagonistic coevolution might
generate incompatibility between male gametes and female reproductive tract and/or
differential sperm competitive ability between the two populations (Anderson &
Simmons 2006; Parker 2006). Differences in sexual conflict and selection can cause
varying degrees of reproductive isolation (e.g., Martin & Hosken 2003; Elgee et al.
2010). The pattern of precedence observed in our study is consistent with that
hypothesis, but our results could also be generated if Inwood males had better sperm
longevity during storage within the female and thus, there were simply more stored
sperm to compete with recently inseminated heterotypic sperm. However, if this were
the case we would also expect to see the same effect within the homotypic cross as
well, i.e., within I x I. Likewise, if recently inseminated Inwood sperm was intrinsically
superior, then we would expect to see the same pattern of last male sperm precedence
in the S x I cross as the I x I cross. But perhaps Snake Island males have sperm with
inferior longevity. Further, it may be that the density and OSR at which the matings
were conducted was perceived differently by males of each population. We are unable
to distinguish among potential mechanisms with these data and a series of experiments
that control sperm numbers across different temporal scales between matings would do
much to shed light on the questions generated by this data set.
Asymmetry in large-male postcopulatory advantage
Perhaps the most surprising result was that body size affected the probability of
paternity for Snake Island males, but only when they mated with females from their
35
own population. This trend was not seen in the three other crosses. There are many
possible mechanisms that might explain these results: Larger Snake Island males may
produce more sperm or have some other advantage in postcopulatory selection;
females may control sperm transfer and preferentially allow more sperm from larger
males; or Snake Island males may tailor their ejaculate to suit their partner (size or
population). Regardless of the mechanism, because the size advantage was nonexistent
when Snake Island males were mated with Inwood females it is suggestive of intersexual
coevolution within the Snake Island population or the lack of the same coevolutionary
selective regime in the Inwood population. Again, experiments investigating these
possibilities would be informative. For example, one could count sperm in the ejaculates
of Snake Island and Inwood males to compare relative sperm numbers per ejaculate and
if they covary with male size differently between populations.
The asymmetry in reproductive isolation we found may be due to fundamental
differences in the mating system ecology between the two populations. In large malebiased mating aggregations sexual conflict via sexual harassment, direct coercive
matings and/or limited precopulatory female choice may be driving the coevolution of
female postcopulatory mechanisms to make the best of a bad situation. If these
patterns are consistent across generations and whatever selective regimes are
consistent gametic isolation may further isolate these populations. Vernal double
matings between these populations while recording the latency to mate would shed
more light on the degree of gametic isolation.
36
SPERM COMPETITION AND MATE-ORDER EFFECTS IN RED-SIDED
GARTER SNAKES
Christopher R. Friesen, Amelia R. Kerns, and Robert T. Mason
Abstract
Promiscuous females may use mate order effects to bias paternity toward fitter males.
In Manitoba, Canada, a female red-sided garter snake has limited precopulatory matechoice within the large courtship aggregations she encounters at spring emergence.
Once mated, a female remains attractive enough to receive courtship and will
potentially remate in smaller aggregations away from the den where larger males have
a mating advantage. If a female uses mate-order effects to ensure that a larger (fitter)
male will sire her offspring, then that second male should have precedence in sperm
competition. We tested for second-male sperm precedence using singly-mated females
that were then assisted to mate with a second male. Average proportion of paternity
was shared equally among the first (P1 proportion of paternity of the first male to mate)
and second males (P2) to mate, and stored sperm (Pss). This may be a case where last
male precedence breaks down with more than two males. All females were spring
virgins, thus sperm stored from fall matings may be important in this system. As the
interval between matings increased P1 increased at the expense of Pss. As the second
male’s copulation duration increased, P1 also increased at the expense of P2. This last
result may indicate female influence over sperm transfer during coerced matings.
37
2.1
Introduction
Female sexual promiscuity is rampant across animal taxa (Smith 1984; Birkhead and
Moller 1998; Simmons 2001). When females are sexually promiscuous within a
reproductive season, sperm competition and cryptic female choice must be considered
important factors in the evolution of mating systems (Birkhead and Pizzari 2002; Shuster
and Wade 2003; Andersson and Simmons 2006). Sperm competition can decrease
variance in fertilization success between males by reducing male fecundity gains from
multiple mating (Lorch 2002). On the other hand, high variance in male spermcompetitiveness and cryptic female choice may increase variance in male fertilization
success and increase strength of sexual selection by reinforcing precopulatory selection
(Eberhard 1996; Simmons 2005). Through cryptic mechanisms, females may bias
paternity toward or away from particular male phenotypes (e.g., Pilastro et al. 2004).
Through their effects on mating systems, postcopulatory sexual selection and sexual
conflict can affect speciation rates by affecting sexual isolation (Coyne and Orr 2004;
Gavrilets and Hayashi 2005). For example, polyandrous insects clades exhibit higher
species richness that is correlated with increased rates of sexual conflict (Arnqvist and
Nilsson 2000; Arnqvist et al. 2000).
Although accumulated evidence confirms the ubiquity and strength of postcopulatory
sexual selection (Birkhead and Pizzari 2002; Andersson and Simmons 2006), this
evidence has come mostly from insects and birds (Wigby and Chapman 2004; Andersson
& Simmons 2006; Birkhead et al. 2008). The paucity of studies reptiles is unfortunate
38
and because they may lead to important insights within the field of postcopulatory
sexual selection due to the prevalence of long-term female sperm-storage (months to
years in some species, Uller and Olsson 2008). Little progress has been made elucidating
postcopulatory sexual selection in reptiles beyond documenting extensive multiple
paternity within the group (Uller and Olsson 2008). An increase in the number of
potential fathers engaged in competition when stored sperm are present may affect
patterns of sperm precedence. Patterns of sperm precedence may no longer exhibit
consistent mate-order effects, which may have implications for evolutionary trajectories
(e.g., Zeh and Zeh 1994). In addition, it is unclear how well sperm stored within the
female’s reproductive tract for long periods can compete with more recently
inseminated sperm.
A simple, yet fundamental, principle of sperm competition is that males increase their
odds of fertilization success by increasing the number of sperm present relative to
competitors within the female at the time of ovulation (Parker 1990). The order in which
two males mate can affect the relative numbers of sperm present at the time of
ovulation and the timing of those matings relative to ovulation can affect paternity
(Birkhead 1998). The female reproductive tract is inhospitable to spermatozoa (Poiani
2006), which are short-lived in most taxa (Birkhead and Møller 1998). Thus, the sperm
of the last male to mate are usually at a numerical advantage due to attrition of his
competitors’ sperm within the female. Assuming females don’t bias paternity (i.e.,
cryptic female choice, Eberhard 1996; Pitnick and Brown 2000), the result is last male
39
precedence: the mating order effect wherein the last male to mate fathers the
preponderance of offspring produced by the female.
Mate-order effects and sperm competition are important in the evolution of mating
systems (Shuster and Wade 2003). Females may evolve to be more selective if courtship
and mating are costly and if first male precedence is the norm (Kokko and Mappes 2005;
Bleu et al. 2012). If last male precedence is the norm, a female may mate
indiscriminately first and then become choosier on successive matings (e.g., Gabor and
Halliday 1997). In other words, females may trade-up over previous mates if there is last
male precedence (Pitcher et al.2003). Further, females may hedge their bets or relieve
conflict by mating with multiple males (Yasui 2001; Simmons 2005). This behavior might
be especially important in mating systems in which females have limited precopulatory
choice (Gowaty 1997). Therefore, after establishing the occurrence of multiple
paternity in a species or population, the next priority should be to characterize the
effects of mating order on paternity if we wish to fully appreciate the mating system
(Shuster and Wade 2003). However, there is one complicating factor in the assessment
of mate-order effects which occurs when females store sperm for long periods before
ovulation. In some taxa, such as honey bees (Apis mellifera), sperm survive for years
within the female reproductive tract (Collins et al. 2004). In reptiles, sperm stored over
long periods are capable of fertilization and have been demonstrated to account for
multiple paternity in captive snakes and turtles (Agkistrodon contortrix, Schuett and
Gillingham 1986; Chrysemys picta Pearse et al. 2001) and one wild population of garter
40
snakes (Thamnophis sirtalis parietalis, Friesen et al. in prep.). Due to a dearth of studies,
the evolution and maintenance of long-term sperm storage and its effect of mating
system evolution in snakes is not understood (Uller and Olsson 2008).
We used a well-studied reptile model, the common garter snake (Thamnophis sirtalis),
which is known to exhibit long-term sperm storage (e.g., Blanchard and Blanchard 1943;
Stewart 1972) to address within-season mate-order effects. Multiple paternity is
common in garter snakes (Blanchard and Blanchard 1941; Gibson and Falls 1975;
Schwartz, McCracken et al. 1989; McCracken, Burghardt et al. 1999; King et al. 2000;
Garner, Pearman et al. 2004; Garner and Larsen 2005; Wusterbarth et al. 2010), and
thus T.sirtalis is an ideal candidate in which to study long-term female sperm storage
and mate-order effects. Previously, we have demonstrated that sperm stored prior to
winter hibernation can compete with sperm from a single vernal mating in T. sirtalis
parietalis (Friesen et al. in prep.). In this report, we investigate the mate-order effects of
two vernal matings and whether stored sperm gain a share of paternity.
41
2.2
Methods
Model system
Red-sided garter snakes of the Interlake region of Manitoba, Canada emerge en masse
from communal limestone hibernacula (dens) in late April each spring. The males form
large aggregations around the den. Females emerge into the large aggregations of males
and mate. During mating, males deposit a gelatinous copulatory plug that prevents
sperm leakage and decreases the opportunity for female remating for up to two days
before it fully dissolves (Shine et al. 2000a; Friesen et al. submitted). Within the large
aggregations in the den, male mating success is essentially random with respect to male
size (Shine et al. 2000b) and female precopulatory choice is limited (Shine et al. 2000c).
However, in smaller aggregations, such as those found in the aspen woods surrounding
the Manitoba dens, larger males have a large and significant size advantage (Shine et al.
2000b). As snakes display indeterminate growth (Fitch 1965), large size indicates that a
male has survived many seasons and thus may carry good genes. Given the random
mating within the den, females may trade-up over their previous mate via sperm
competition (in sensu Thornhill 1983, Simmons 1987), if second male precedence is the
norm. We opportunistically utilized females that were mated during an experiment to
measure the energetic costs of courtship using doubly-labeled water (Friesen Powers &
Mason in prep). These females were assisted to mate a second time, and we assigned
paternity to their offspring to assess mate-order effects in this species. Given that males
invest heavily in the production of a copulatory plug, presumably to prevent second
42
matings, we predict that second male precedence should be the rule in this species.
However, if stored sperm are present and the female mates with two more males,
second male precedence may breakdown as Zeh and Zeh (1994) demonstrated in the
Pseudoscorpion, Cordylochernes scorpioides.
Animal collection and mating trials
A day before mating trials commenced, animals were collected by hand from a
population near Inwood, MB and taken to the Chatfield Research Station 16 km away.
They were kept in seminatural nylon arenas and were provided water ad libitum. All
females were seasonal virgins as they were collected immediately upon emergence
from winter hibernation before mating. Thus, they did not have the opportunity to mate
with any males except those they may have mated with prior to hibernation. These
matings would be the source of stored sperm in our paternity analysis, as they would
not be attributed to our known first or second male. Over the course of the next 20 days
(25 April to 15 May, Figure 2.4) a group of 24 randomly selected males were allowed to
court and mate with these females in 1mx1mx1m seminatural arenas on each day warm
enough to allow vigorous courtship. When a pair had mated, we placed the pair in a
smaller arena where they were under constant observation so we could time copulation
duration (± 10s). Each male mated an average of 2.16 times (0-5 matings per male). The
first male’s number of matings did not affect the likelihood of a female giving birth
(Fisher’s exact test, P = 0.424). After the initial mating, the females were kept separate
from males until 15 May, when each female was assisted to mate with a second male. In
43
this way, we were able control the timing of the last mating relative to ovulation (late
June in this population, Garstka et al. 1982). The males were collected from the same
den site near Inwood on 14, May, the day before the last mating. Only 20% of females
will remate without assistance in arenas trials (Shine et al. 2000a). Thus, to increase the
number of doubly-mated females we assisted remating by gaping the female’s cloaca as
they experienced courtship within the arenas. As a courting male aligned with the
female we used a blunt probe to lift the ventral scale that covers the opening to her
cloaca. Typically the courting male would then evert one of his two hemipenes into the
female’s cloaca in only a few seconds. Copulation duration was measured as previously
described. Over all matings there was no difference in copulation duration between a
female’s first and second mating (Signed rank test W = 235, P = 0.244). The assisted
mating method allowed us to conduct all of the second matings on a single day when
weather permitted. Each male’s size (mass ± 0.1 g, and snout to vent length (SVL) ±
1mm) was recorded. Tail tips were collected from each male for genotyping (Garner,
Pearman et al. 2004). Female size (svl and mass) was recorded after the second mating.
Females were returned to laboratory facilities at Oregon State University where gravid
females were either kept alone or with a non-gravid female until they gave birth in late
summer, thus maternity was certain. Of the 52 doubly-mated females 26 (50%) gave
birth which is a typical parturition rate for capital breeders that rely on stored energy to
determine their reproductive state (Gregory 2006). Litter sizes ranged from four to
thirty one ( ̅ ± SEM; 15.73 ± 1.26), with a total of 409 offspring produced.
44
Molecular methods
DNA was extracted from tail tip tissue (Garner et al. 2004) in 200 µl 5% chelex 1% ProK
solution incubated at 56°C for 2 hours and then 8 minutes at 100°C. We used three
microsatellite loci to exclude the focal male from paternity: Ts1 (McCracken et al. 1999)
and Nsµ2 and Nsµ3 (Prosser et al. 1999). All three loci were multiplexed in a single 12 µl
PCR reaction [1 µl template; 6.25 µl Multiplex mix (Qiagen cat no. 206143); 1.25 µl 2 µM
of each primer (Invitrogen); 3.5 µl molecular grade H20]. PCR reactions were carried out
in a Biorad thermal cycler (C1000). Amplification conditions consisted of 15 min at 95°C,
followed by 35 cycles of 30 sec at 95°C and 1:30 min at 58°C, and 1:30 min at 72°C
followed by a 10 min at 72°C elongation phase. Reaction products were analyzed in an
ABI 3100 genetic analyzer and the alleles were visualized using ABI Genotyper software.
Peaks were assigned manually and each offspring was checked against the mother’s
genotype. Any offspring that did not have a maternal allele was rerun along with the
mother’s sample and a random subset of siblings to check for errors. In total 26 families,
and all 409 offspring, were successfully and satisfactorily genotyped with a maternal
allele.
We used an exclusion based protocol for paternity assignment. We estimated average
exclusion probability using CERVUS 3.0.3. (Kalinowski et al. 2007). The three loci chosen
for this study were all highly polymorphic and loci Ts1 and Nsµ3 were in HWE (Ts1, χ2 =
298.9, d.f. 325, P = 0.847; and Nsµ3, χ2 = 50.386, d.f. 66, P = 0.923), however, locus
Nsµ2 was not (Nsµ2, χ2 = 503.5, d.f. 190, P < 0.001). The average exclusion
45
probabilities, based on the genotypes from 56 random adults from a previous study
were: Ts1 (1 – 0.1160) = 0.884 exclusion probability (second parent); there were 26
alleles found in 56 genotyped individuals; Observed heterozygosity, H0 =0.9643. Nsµ2 (1
- 0.1654) = 0.8346 exclusion probability (second parent): there were 20 alleles found in
56 individuals, H0 = 0.5357. Nsµ3 (1 - 0.2508) = 0.7492 exclusion probability (second
pair); there were 12 alleles found in 56 individuals, H0 = 0.8929 (Friesen et al.
submitted). The combined average exclusion probability using all three alleles is (10.0048) = 0.9952 and any pair of loci yields greater than 0.95 confidence of correctly
excluding our focal male. Analysis of the adults within this study established that the
single locus exclusion probability of locus Ts1 was 0.98, provided the maternal and focal
male’s genotypes did not match. Thus, we relied on this locus alone when the other loci
were uninformative. A conservative estimate of the minimum number of fathers per
litter was calculated by dividing the number of paternal alleles by 2 after excluding
alleles assigned to the known males. We then added the known males to estimate the
total minimum number of fathers in a litter.
46
2.3
Results
Distribution of paternity and mate-order
We detected an average of 2.62 fathers per litter (1-4 fathers, SEM ± 0.170), Figure 1.1.
On average, the first male to mate (P1) fathered 33% ( ̅ ± SEM; 0.33 ± 0.063) of the
offspring across litters. Larger litters were not likely to have more fathers than smaller
litters (SLR, F1, 24 = 0.664, P = 0.424). The second male to mate (P2) fathered 34% (0.34 ±
0.064) of the offspring. We could not assign paternity to 33% (0.33 ± 0.058) of the
offspring, thus they were attributed to stored sperm (P ss) from autumnal or previous
vernal matings. There was no difference in average paternity attributable to male mateorder effects (Kruskal–Wallis one-way ANOVA on ranks, H = 0.103, df = 2, P = 0.950),
Figure 1. 2. There was no effect of male size or female size on paternity. There was no
effect of the minimum number of fathers on P1 (F1, 24 = 0.333, P = 0.333), P2, (F1, 24 =
0.029, P = 0.865) or Pss (F1, 24 = 0.931, P = 0.345).
14
12
Count
10
8
Figure 2.1: Minimum number of
fathers per litter. We detected an
average of 2.62 fathers per litter
(1-4 fathers, SEM ± 0.170)
6
4
2
0
1
2
3
Minimum number of fathers
4
47
Proportion of offspring fathered
1.0
0.8
Figure 2.2: Proportion of paternity
attributed to stored sperm (Pss), the
first male to mate (P1), and the last
male to mate (P2). The boxes enclose
50% of the data, the whiskers are 1.5
interquartile ranges, the solid line is
the median, and the dashed line is
the mean.
0.6
0.4
0.2
0.0
Pss
P1
P2
Source of paterity (male mate order)
Sperm depletion effects on paternity
As noted above, most of the first males to mate had mated multiply. Seven pairs of
females (N = 14) that gave birth were used to assess the effect of the number of matings
a first male had on his male fertilization success (N = 7 males, i.e., P1). Because each
female within in a pair mated with the same male first and then mated a second,
different male, we could assess the effect of the first male’s ability to defend paternity
after he had mated more than once using repeated measures. The number of times a
male mated (male mate number) did not have a significant effect on paternity of the
first male to mate (P1) after successive matings (Paired t-test; tdf=7 =0.727, P = 0.491),
Figure 2.3.
48
1.0
Proportion of paternity
0.8
0.6
Figure 2.3: Before and after line plot
showing the effect of male matenumber on fertilization effects; first
mating (a), and a male’s second or third
mating (b). (Paired t-test; tdf=7 =0.727, P
= 0.491)
0.4
0.2
0.0
a
b
Successive male mating
Effect of the interval between matings and the 2nd male’s copulation duration on paternity
We conducted backward stepwise regression model selection (BSRMS) on arcsine(aqrt)
transformed proportions of offspring fathered, which revealed that the second male’s
copulation duration (CD2) and mating interval (MI) were significant predictors of P 1 (P1 ~
CD2 + MI; Adj. R2 = 0.492, F2, 24 = 12.606, P < 0.001). BSRMS selected the second male’s
copulation duration as the only significant predictor of P2 (P2 ~ CD2; Adj. R2 = 0.258, F1,21
= 8.646, P = 0.008). BSRMS found that a male’s mate-number (MM#) affected the Pss (Pss
~ MM#; Adj. R2 = 0.135, F1,21 = 8.646, P = 0.048). However, this relationship was driven
by a single male’s fifth mating in which all the offspring were attributed to stored sperm.
If this observation is removed from the analysis, then male mate-number is not
significant (Adj. R2 = 0.000, F1,21 = 0.392, P = 0.537), Figure 2.4.
49
a
b
c
Figure 2.4: Individual regressions of proportion of offspring sired. Each panel represents source of
paternity Pss , P1, and P2 left to right. 4a. the effect of the second male to mate’s copulation duration on
paternity. 4b. the effect of the interval between the first and second matings. Because all the second
th
matings occurred on the same day (May, 5 ), longer intervals mean that the first matings occurred earlier
in the season. Figure 4c., the timing of matings relative to ovulation in this species. Thus, it is useful to
think of the interval between matings as also representing the time between the first male’s mating and
ovulation. All sperm were stored in the female for >40 days before ovulation in late June (Whittier and
Crews 1986).
50
2.4
Discussion
Multiple paternity has been established in five of six species within the genus
Thamnophis that have been examined thus far (Wusterbarth et al. 2010), and 19 of 34
(55.9%) litters genotyped in T. sirtalis exhibited multiple paternity (Uller and Olsson
2008). However, few studies have investigated mate-order effects in reptiles and only
two in snakes; Hӧggren and Tegelstrӧm (2002), found evidence for first-male advantage
in eight litters of captive Vipera berus, an Eurasian viper. In a previous study of red-sided
garter snakes from this population, we found 85% of litters exhibited multiple paternity
and that sperm from spring matings had precedence (Friesen et al. in prep.), and this
would represent last male advantage. However, in contrast with Hӧggren and
Tegelstrӧm’s work (2002), the competing sperm in our study were presumably at a
disadvantage as they were stored within females across winter hibernation and would
have suffered attrition. Nevertheless, the stored sperm accounted for a significant
proportion of offspring across litters ( ̅ = 25%). This study confirms our previous result
in that sperm stored over winter are an important source of paternity ( ̅ = 33%).
Further, average paternity was shared equally among the first (P1) and second male (P2)
to mate, and stored sperm (Pss). Almost 60% of the litters had more than two sires, this
result is consistent with other studies of multiple paternity in Thamnophis in which two
to three sires were common (McCracken et al. 1999; King et al. 2000; Garner and Larsen
2005). Contrary to our previous work in which females had only one vernal mating,
variance in fertilization success was extremely high. Male size did not explain any of the
51
variation in paternity, nor did a male’s number of matings, although the small sample
size may have precluded detecting the trend. Zeh and Zeh (1994) found that strong last
male precedence decreased and became more variable when they altered the interval
between matings in the harlequin-beetle-riding pseudoscorpion. Further, they found an
even distribution of paternity among males when females were mated with a third male
(Zeh and Zeh 1994). As with the Zeh and Zeh study, our analyses have uncovered two
significant sources of variation in paternity observed by these authors: a) the interval
between mating, and unexpectedly, b) a negative association between the second male
to mate’s copulation duration and his paternity.
The interval between matings is easiest to interpret. Given that all the second matings
took place at the same time, longer intervals can be more clearly interpreted as the first
mating having taken place earlier in the season. The first male’s paternity increased the
earlier he mated in the season. The gain in the first male’s paternity seems to come at
the expense of paternity assigned to stored sperm. The second male’s paternity was
unaffected by the interval between matings. This result is mostly likely because all
second matings occurred on the same day relative to ovulation. We suggest that sperm
from the first matings that occur earlier in the season have more time to displace or
overlay sperm stored prior to hibernation. Further, sperm storage tubules (SSTs) have
been described in garter snakes as invaginations in the anterior infundibulum of the
oviduct, and sperm tend to fill the more posterior SSTs first (Fox 1956). Devine (1984)
suggested that second male precedence should be the rule if the anterior SSTs were
52
filled by sperm from successive inseminations and were thus more proximate to ova at
the time of ovulation. Given the paucity of literature on mate-order effects in snakes,
this study is useful despite being opportunistic. Mating trials conducted with variable
mating intervals conducted both early and late in the breeding season would go a long
way toward clarifying the effect of mating interval on patterns of precedence.
It may seem paradoxical that longer copulations by the second male reduced his
paternity. Indeed, longer copulation durations are positively correlated with the amount
of sperm transferred in at least two species of lizard (Tokarz 1999, Olsson 2001), as well
as many other taxa (Birkhead and Møller 1998; Simmons 2001). However, sperm
numbers are unaffected by copulation duration in this population of garter snakes
(Friesen et al. in prep.). Further, as P2 decreased, P1 increased, and with no effect on Pss.
There are a few intriguing hypotheses to explain this pattern: 1) the muscular
contractions or twisting of the oviduct (Andrén and Nilson 1982; Siegel and Sever 2006)
caused by the second male’s insemination aided transport of the first male’s sperm, 2)
the second male’s ejaculate physically displaced the first male’s sperm further into the
oviduct, 3) the female may have a mechanism to reduce the paternity of a coercive male
such as constriction of the vaginal pouch (references in Siegel and Sever 2006). We
suspect the third explanation is most likely because the second insemination was
assisted, not voluntary, and thus it could be considered a simulated coerced mating.
Increased copulation duration could indicate female resistance to sperm transfer. In
addition, subsequent work has found that there is an oviductal sphincter that may
53
prevent sperm from entering the oviducts during copulation (Friesen et al. in prep.), and
the vaginal pouch is highly muscularized and may also limit sperm and or plug transfer.
The plug material functions to prevent leakage (Friesen et al. in prep.). Given the
random mating within the den, and lack of precopulatory choice, it might behoove the
female to reduce sperm transfer during copulation and thus, we would not find
increased sperm insemination with increased copulation duration. This hypothesis is not
unfounded. In jungle fowl, females are able to eject the sperm of an unwanted suitor of
low social rank (Pizzari and Birkhead 2000), and female dunnocks eject sperm in
response to cloaca-pecking by their dominant mate (Davies 1983). The vaginal pouch of
T. s. parietalis is thickened with muscle, so a similar mechanism may exist in snakes as
that in birds. None of these hypotheses are mutually exclusive and the second male to
mate may unwittingly aid his rival.
These results invite additional experimentation to unravel the many questions we have
exposed. For example, one could systematically manipulate mating interval to test its
effect on mate order. In addition, this experiment should be duplicated early and late in
the breeding season to account for the transport time available for sperm to reach
fertilization sites and perhaps displace stored sperm. Comparisons of paternity patterns,
sperm counts, plug masses, and copulation durations between assisted matings and
natural single vernal matings may reveal whether females exert control of sperm
transfer during coerced matings, and could also be accompanied by experiments that
aim to test the relationship between sperm transfer and copulatory plug deposition,
54
which is also a potential source of sexual conflict. Because stored sperm can be a
substantial source of paternity in this system, long-term sperm viability and longevity
may be an important target of selection.
55
NOT JUST A CHASTITY BELT: THE FUNCTIONAL SIGNIFICANCE OF
MATING PLUGS IN GARTER SNAKES REVISITED.
Christopher R. Friesen, Richard Shine, Randolph W. Krohmer and Robert T. Mason
Abstract
During the spring emergence of red-sided garter snakes (Thamnophis sirtalis parietalis) in
Manitoba, Canada, the operational sex ratio is strongly skewed toward males who scramble to
locate and court newly emerged females. A high frequency of multiple paternity of litters
suggests that the females are promiscuous; the gelatinous copulatory plugs (CPs) deposited by
males may confer fitness benefits via passive mate-guarding . Because precopulatory female
choice is limited in large mating aggregations, sexual conflict may place a premium on
preventing females from ejecting male sperm. In snakes, sperm are produced in the testes and
delivered through the ductus deferens, and the CP is thought to be produced by the renal sexual
segment and conveyed through the ureter. We manipulated the delivery of the two fluids
separately by surgically ligating the ducts. Ureter-ligated males did not produce a CP, causing
their sperm to leak out of the female’s cloaca immediately after copulation. Contrary to
previous suggestions, histology revealed many sperm distributed throughout the CP. Thus, the
CP may function as a spermatophore: the protein matrix contains sperm which are liberated
gradually as the plug dissolves. The likelihood of a male depositing a CP fell significantly after his
second mating, perhaps limiting his reproductive success. Finally, genotyping of the litter from a
female that mated while she retained the CP of her first mating indicates that the CP does not
assure the first male’s paternity. These results challenge the hypothesis that passive mateguarding is the primary function of the garter snake CP.
56
3.1
Introduction
Female sexual promiscuity is phylogenetically widespread (Smith 1984; Birkhead &
Møller 1998). When a female mates with multiple males, the sperm of these males may
commingle within her reproductive tract and compete for the fertilization of her ova, a
phenomenon known as sperm competition (Parker 1970). Males can gain an advantage
in sperm competition by increasing the number of sperm they inseminate relative to
their rivals (Parker 1990). However, an alternative strategy is to reduce the risk of sperm
competition altogether by preventing a mate from remating with another male. Males
can limit remating opportunities by guarding their mates and/or by prolonging
copulations and, thus, reducing the risk of sperm competition (Birkhead & Møller 1998;
Simmons 2001). A male may also passively guard a female by depositing substances that
occlude the opening to her reproductive tract (known as copulatory, mating or vaginal
plugs) (Birkhead & Møller 1998; Simmons 2001). Just as female sexual promiscuity is
taxonomically widespread, so too are copulatory plugs (henceforth, CPs: Voss 1979;
Birkhead & Møller 1998; Shine, Olsson & Mason. 2000; Simmons 2001; Jia, Jian & Wang
2002; Poiani 2006; Althaus et al. 2010).
Although the passive mate-guarding or “chastity-enforcement” hypothesis has received
some support (e.g., Shine et al. 2000; Simmons 2001), the efficacy of mating plugs as a
paternity guard has also been called into question (e.g., Michener 1984; Moreira &
Birkhead 2004). This controversy has caused researchers to re-visit alternative
explanations for the evolution of copulatory plugs. Voss (1979) succinctly summarized
57
several of these alternative hypotheses for plug function: “1) permit a gradual release of
spermatozoa within the female tract as they disintegrate (Asdell 1946), 2) prevent
leakage of spermatozoa from the vagina (Leuckart 1847), 3) induce pseudopregnancy in
the female (Long [and Evans 1922]), 4) transport sperm through the cervix (Blandau
1945), 5) prevent subsequent insemination of females by other males (Martan &
Shepherd 1976), or 6) act as an antiaphrodisiac (Happ 1969).” Tests of these alternative
hypotheses are sparse. Further, these functional hypotheses are not mutually exclusive.
For example, in scorpions the CP performs both mate guarding and sperm storage
functions (Althaus et al. 2010), and in ground beetles the mate guarding and
antiaphrodisiac hypotheses have found support (Takami et al. 2008). In some mammals
the passive mate guarding hypothesis has been rejected in favor of the sperm leakage
hypothesis, because paternity was not assured by the first male to mate (ground
squirrels, Michener 1984 and Koprowski 1992; deer mice, Dewsbury 1988; civets, Jia et
al. 2002). However, sperm leakage in the absence of plug deposition was not
experimentally tested in these studies, and confirmation of one hypothesis does not
falsify or support the others. Each functional hypothesis needs to be tested
independently.
One species in which multiple hypotheses of plug function have been experimentally
evaluated is the red-sided garter snake (Thamnophis sirtalis parietalis). The CP of the
red-sided garter snake is the largest plug among reptiles (Olsson & Madsen 1998), and
can represent more than 0.4% of a male’s mass (Shine et al. 2000; Friesen & Mason in
58
prep.); this is the equivalent of a 90 kg human male producing a 360 ml ejaculate (based
on a density of 1g/ml for human semen; Matson et al. 2010). This is a substantial
energetic investment (Friesen, Powers, & Mason in prep.), and given such an enormous
investment, one would expect a CP to provide significant fitness benefits to the male
who produces it (e.g., Dewsbury 1982).
Red-sided garter snakes of Manitoba (in Canada’s Interlake region) emerge en masse
each spring from communal hibernacula (Gregory 1974). Males emerge first and remain
around the den for 2-4 weeks before migrating to summer feeding grounds, while
females emerge throughout the 4-6 weeks of spring emergence, but stay an average of
four days before migrating (Gregory 1974; Shine et al. 2001, 2006a). This pattern of
emergence generates an operational sex ratio (OSR) that is highly skewed towards
males, generating intense competition among males to scramble for mates (Emlen &
Oring 1977). Mating aggregations can contain up to 69 males to one female, with an
average of four males per group and with 80% of groups containing 10 or fewer males
(Shine et al. 2001). Two hypotheses for copulatory plug function have been explored in
this system: the antiaphrodisiac and the mate guarding hypotheses. Both of these
hypotheses propose that the CP functions to reduce the likelihood of rematings by
females.
(1) Antiaphrodisiac hypothesis: Mated female plains garter snakes (Thamnophis radix)
and T. sirtalis with CPs are less sexually attractive and receptive than are unmated
59
females, suggesting that the plug may discourage courtship (i.e., have an antiaphrodisiac
function: T. radix Ross & Crews 1977, 1978; T. sirtalis Devine 1977; Whittier, Mason &
Crews 1985; Shine et al. 2000). However, this reduction in female receptivity may be
due to the presence of prostaglandins in the seminal fluid rather than the plug itself
(Whittier & Crews 1986; Shine et al. 2000a).
(2) Chastity belt hypothesis: Supporting earlier speculation by Devine (1975), Shine et al.
(2000a) found support for the hypothesis that the garter snake CP acts to discourage
remating by females. Mated females that had the plugs removed shortly after mating (<
5 min.) remated at a higher rate than did females with plugs removed a few hours after
mating, or females in which the plug dissolved naturally after about two days. These
results were interpreted as support for the mate guarding hypothesis, but it seems that
the possible effect on female receptivity was not controlled for.
Unfortunately for the “chastity belt” hypothesis, some inconvenient observations raise
questions about the plug functioning as a form of mate guarding even in this wellstudied example. First, some females are able to remate even while plugs occlude their
cloacas (Shine et al. 2000a; Whittier & Crews 1986; Friesen personal obs.). Importantly,
the plugs form only a temporary barrier to subsequent mating. Although most mating in
this population occurs at the den during the spring emergence, the garter snake
breeding season is protracted and much longer in duration then a copulatory plug
remains in a female’s reproductive tract. The copulatory plug does not even last the
60
average duration that a female remains at the den where males are constantly courting
her: it only takes 2-3 days for the plug to dissolve completely and the plug can be easily
displaced after only 1.5 days if the temperature is high enough (Shine et al. 2000a). On
average, female garter snakes leave the den to migrate to feeding grounds about four
days after they emerge (Shine et al. 2001; Shine et al. 2006a). Thus, there is abundant
opportunity for females to remate before migrating. Reinforcing the ineffectiveness of
the copulatory plug as a chastity belt, multiple paternity of litters is ubiquitous in the
genus Thamnophis (Schwartz, McCracken & Burghardt 1989; McCracken, Burghardt &
Houts 1999; Garner et al. 2002, 2004; Wusterbrath et al. 2010). Although long-term
sperm storage may account for some multiple paternity (Uller & Olsson 2008; Uller,
Stewart-Fox & Olsson 2010), it is clear that the plug does not necessarily prevent
multiple mating.
Many alterative hypotheses for the copulatory plug remain to be explored in this
system. For example, the copulatory plug may function to keep a male’s sperm in the
female reproductive tract. If sperm leaks from the female after mating in the absence of
plug deposition, then mate guarding, antiaphrodisiac and the reduction in female
attractivity are secondary functions. It is possible that sperm move up the reproductive
tract during copulation and do not leak from the female’s reproductive tract. In this case
the previously supported hypotheses, such as mate guarding, fully explain the adaptive
significance of copulatory plugs in this system (e.g., Shine et al. 2000a). Unfortunately,
the fate of sperm is rarely directly examined in studies of copulatory plug function. In
61
garter snakes, preliminary observations found abundant sperm in the oviducts of
recently mated females (Devine 1975). If this were the case more generally, then the
sperm leakage hypothesis would seem less likely. However, in preliminary studies aimed
at the collection of sperm to evaluate variation in ejaculate quality, we did not find
sperm in the oviducts (Friesen & Mason pers. obs.), supporting the hypothesis that the
primary function of the plug is to reduce or eliminate sperm leakage.
To test the hypothesis of sperm leakage, one must be able to prevent a plug from being
deposited while allowing males to transfer sperm as usual. Squamate sauropsids are
uniquely suited to this type of study because the sperm and plug are delivered via
completely different tracts. In many squamates (lizards, snakes and worm lizards), the
distal portion of the nephron preceding the collection ducts of the kidneys is seasonally
hypertrophied or regressed in males, but not females (in which it is never
hypertrophied: reviewed in Krohmer 2004). This sexual dimorphism suggests that this
portion of the kidney may serve a sexual function; hence this portion of the kidney has
come to be called the renal sexual segment (RSS) (Burtner, Floyd & Longley 1965).
Seasonal regression and recrudescence of the RSS is also influenced by testosterone,
which further supports its sexual function in males but not in females (Bishop 1959;
Krohmer 2004). Volsøe (1944) proposed that the copulatory plug was produced by the
RSS. Given these facts, Olsson & Madsen (1998) suggested that squamate sauropsids
would provide a perfect model to test the fate of sperm in the absence of plug
formation, because the sperm and plug are delivered separately via two different duct
62
systems and could be manipulated separately. However, the true function of the RSS
was still in doubt (Krohmer 2004).
To experimentally test whether the copulatory plug prevents sperm leakage, we first
had to determine the RSS is in fact the source of the copulatory plug material. In this
experiment, we observed the fate of sperm in females in which plugs were not
deposited. We also examined the distribution of sperm within plugs that were removed
immediately after copulation was terminated. Further, we report data on the incidence
of plug deposition in relation to the number of matings a male obtains to determine if
males are limited in the number of plugs they can produce (Shine et al 2000a). This
question addresses the ecological relevance of plug function, because if plugs are always
produced, then the prevention of sperm leakage is less likely to be the primary target of
selection. In this last analysis, we also assessed the effect of copulation duration on
plug deposition. Finally, we tested the effectiveness of the plug as a barrier to a second
male’s sperm. In this case, we assigned paternity to the litter of a previously plugged
female that was caught while mating with a second male.
63
3.2
Methods
Twenty-nine male red-sided garter snakes were collected from the Inwood quarry in
Manitoba, Canada (Inwood den: 50° 31.58’N 97°29.71’W) on May 10, 2008. They were
placed in 1m x 1m x 1m seminatural enclosures located at the Chatfield Research
Station until surgeries were performed. Water was provided ad libitum, but food was
not (snakes are aphagous at this time: O’Donnell, Shine & Mason 2004). Body mass and
snout-to-vent length (SVL) were recorded.
Surgeries
The reproductive tract of male squamates (snakes, lizards and worm lizards) is paired.
The ductus deferens (from the testes) and the ureters (from the kidneys) do not
converge until they reach the posterior section of the cloaca. As a result, the ureters and
ductus deferens can be manipulated independently, thus separating the contributions
of each to the final ejaculate.
Twenty-nine males were randomly assigned to one of three treatments: ureter ligation
(Ux), vasectomy (Vx) or sham surgery (N = 5 Ux males; N = 12 Sham; N = 12 Vx). For
anesthesia, 15 mg/kg (0.003 ml of 0.5% methohexital sodium/g body mass) was
administered subcutaneously at the juncture between the dorsal and ventral scales,
approximately 20 cm from the head of the snake (Preston, Mosley & Mason 2010). An
incision was made in the skin between the first and second lateral scale rows directly
dorsal to the fifth ventral scale (counting from the cloacal scale). The surgical window
traversed no more than four scales, and the incision was made from posterior to
64
anterior with corneoscleral scissors that were sterilized in a dry-bead sterilizer at 260ºC
for >15 sec. The peritoneal membrane was then cut using the same scissors and made
wider than the skin window to allow for easier access to the ureter. The ureters (Ux
males) or ductus deferens (Vx males) were isolated from the viscera and ligated using
sterile suture (4-0 G silk). The surgical window was then sutured using sterile 4-0 G silk.
In sham surgeries , the mesentery enclosing the ureters and ductus deferens was
isolated and exposed but not ligated. Subsequently, the incision site was rinsed with
Ringer’s solution (reptile physiological saline), then swabbed with 100% ethanol. The
snakes were then placed in an aquarium with a heating pad (30ºC) and monitored every
10 minutes for the first hour after surgery until return of the righting reflex. Males of
this species regain their drive to court very soon after surgery and show no deficit in this
ability (e.g., Nelson, Mason & Krohmer 1987). All animals survived surgery and engaged
in courtship within 24 hours after recovery.
Mating trials
The males were allowed to court and mate with females in small circular arenas (45 cm
dia. x 75 cm tall). Males were randomly assigned to one of three small circular arenas
(45 cm dia. x 75 cm tall) with up to two females at a time. This density of males is
common in and around the dens (Joy & Crews 1985; Shine, Langkidle & Mason 2004).
Each indoor arena was heated by a 250 watt heat-lamp 1 m above the animals. We
measured external body temperatures with a laser thermometer and adjusted the
output of each lamp using dimmer switches to maintain optimal body temperature (29-
65
30 : Kitchell 1969; Hawley & Aleksiuk 1975). Courtship was observed closely so that we
could begin timing copulation duration. One minute after copulation commenced, the
pair was removed to a separate circular arena to prevent interference from other
snakes. Less than 30 seconds after copulation terminated, each female was inspected
for plug deposition. In addition, a sample of fluid from outside the cloaca of females that
mated with Ux males was collected by wiping the area with a slide to check for the
presence of sperm. For comparison, we also checked five naturally plugged females for
the presence of sperm outside the cloaca using the same method. The slides were
viewed with a compound microscope (10x objective) and photos were taken to
document sperm leakage (Figure 1). The Ux males were euthanized and the ureters
were inspected to ensure that the ligation was effective and to note the condition of the
ureters after copulation. All ligations appeared effective and we observed what was
apparently plug-material accumulated within the ureters anterior to the ligation.
Condom study
In mid-afternoon (1400 to 1600 h) on 15 May 2004, we allowed 14 females to mate in
outdoor enclosures. Seven control females were kept in identical enclosures, but
without males. We checked enclosures regularly for matings, and within 10 minutes of
the cessation of copulation, we collected the females (plus equivalent numbers of
control females) and took them to an adjacent field laboratory where we manually
removed the recently-deposited mating plugs from half of the mated females. The plugs
were placed in 70% alcohol for later analysis of sperm numbers. On all females, we used
66
tape to attach a balloon to the body above the cloaca (i.e., enclosing the cloaca and tail),
such that any materials leaking from the cloaca would be collected in the balloon. The
females were kept overnight in a cage that was large enough (1 X 1 m) to allow them to
move around. The following morning (0800 to 100 h) we removed the balloons, flushed
out their contents, and preserved them in 70% alcohol. Those flushings were later
examined under a haemocytometer to count sperm.
The effect of multiple mating on plug deposition
To estimate the likelihood of plug deposition with successive matings, we analyzed data
collected from matings from the same Vx and Sham males as described above.
Vasectomized (Vx) and Sham males were allowed to mate over the course of six days to
test the effect of multiple mating on plug deposition. The arenas and experimental
design were identical to those described above.
Sectioning of the plug
Our attempts to collect whole ejaculates from the female’s vaginal pouch and oviducts
yielded very few sperm, suggesting that most of the sperm may be contained within the
copulatory plug itself. To address this question, we examined five copulatory plugs
removed from females that mated at the Inwood den, immediately after copulation
terminated. A blunt probe was gently inserted between the plug and the wall of the
cloaca to loosen the plug (Shine et al. 2000a). Plugs were fixed in 4% PFA in 0.1M PBS.
Plugs were dehydrated in progressive alcohol, cleared in toluene and embedded in
paraffin, and cut on a rotary microtome at 10-15 µm thickness. The sections were
67
collected on gelatin-coated slides and allowed to dry. We used two staining methods:
Trichrome (hematoxylin, Biebrich Scarlet-Orange, fast green), and Regressive H&E
(hematoxylin and eosin), which yields blue nuclei and pink cytoplasm but does not stain
proteins. After staining, we dehydrated in progressive alcohols and xylene and covered
the sections with Permount (Fisher) (Figure 3.1).
Figure 3.1: A) Depiction of how the plug was situated in the female’s cloaca, the distribution of sperm
within the plug, and the approximate location of the photographed sections [Figure 1 B–G below
(redrawn after Devine (1975)].
68
69
Figure 3.1 B-G: Cross-sections of a copulatory plug stained with Regressive H & E. The dark stains are
masses of sperm heads and the unstained area within the plug is proteinaceous RSS secretions. The bulk
of the sperm is aggregated, but close inspection reveals that sperm are numerous throughout the plug in
all sections. For example, section one (1-B) is the most caudal and shows masses of sperm to the top and
right side of the picture. The photos of sections 1-5 were all taken using the 4x objective and the scale bar
equals 0.2 mm, and the photo of section six (1-G; the most cranial) was taken using the 10x objective and
the scale bar equals 0.1mm. The total length of sperm cells in this species is 100-110 µm. Photos were
taken with an Olympus DP-5 digital camera mounted on an Olympus CX31 compound microscope with
the filter set to bright field. The images were captured and the scale bars added with Cell-Sense software
from Olympus.
Paternity analysis of doubly-mated female
We collected one female that was found mating with a male at the Inwood den. This
female already contained a copulatory plug, and the second male was depositing a plug
dorsally to the first plug. Once copulation terminated, we collected a tissue sample (tail
tip, e.g., Garner et al. 2002) and a small portion of the first copulatory plug. We also
collected a similar sized portion of the second copulatory plug, to control for the effect
of removing part of the first copulatory plug. DNA was extracted from tail tip tissue in
200 µl 5% chelex/1% ProK solution incubated at 56°C for two hours and then eight
minutes at 100°C. We used three microsatellite loci to exclude the focal male from
paternity: Ts1 (McCracken et al. 1999) and Nsµ2 and Nsµ3 (Prosser, Gibbs &
Weatherhead 1999). All three loci were multiplexed in a single 12 µl PCR reaction [1 µl
template; 6.25 µl Multiplex mix (Qiagen cat no. 206143); 1.25 µl 2 µM of each primer
(Invitrogen); 3.5 µl molecular grade H20]. PCR reactions were carried out in a Biorad
thermal cycler (C1000). Amplification conditions consisted of 15 min at 95°C, followed
by 35 cycles of 30 sec at 95°C and 1.5 min at 58°C, and 1.5 min at 72°C and followed by a
single 30 min extension period at 60°C. Reaction products were run in an ABI 3100
70
genetic analyzer and the alleles were visualized using ABI Genotyper software. Peaks
were called manually and each offspring was checked against the mother’s genotype.
We used an exclusion-based protocol for paternity assignment. We estimated average
exclusion probability in CERVUS 3.0.3. using genotypes from 56 random adults
(Kalinowski, Taper & Marshall 2007). Ts1 (1 – 0.1160) = 0.884 exclusion probability
(second parent); there were 26 alleles found in 56 genotyped individuals, H0 =0.9643.
Nsµ2 (1 - 0.1654) = 0.8346 exclusion probability (second parent); there were 20 alleles
found in 56 individuals, H0 = 0.5357. Nsµ3 (1 - 0.2508) = 0.7492 exclusion probability
(second pair); there were 12 alleles found in 56 individuals, H0 = 0.8929. The combined
average exclusion probability using all three alleles is (1-0.0048) = 0.9952 and any pair of
loci yields greater than 0.95 confidence of correctly excluding our focal male (Friesen et
al. in prep.).
71
3.3
Results
Ureter ligation prevents plug deposition
All five of the Ux males, 10 of 12 Vx, and 10 of 12 Sham males mated. However, none of
the five Ux males deposited plugs compared with 9 of 10 Sham (S) males and 10 of 10 Vx
males. Chi-square tests do not accommodate structural zeros in the contingency table
(i.e., empty cells) so we used a Monte Carlo test of independence between the rows and
the columns in XLSTAT (based on 5000 simulations) which revealed the probability of
plug deposition differed significantly among treatments (χ2 = 20.066, 1 d.f., P = 0.0002).
Logistic regression in XLSTAT (using the ML Newton-Raphson algorithm) confirmed this
result and provided a pairwise comparison of treatments that showed no difference in
plug deposition between Vx and Sham males on their first mating: Model: -2 Log
(Likelihood) χ2 =20.756, P < 0.0001; S vs. Vx, χ2 = 0.456, P = 0.500; S vs. Ux, χ2 = 5.194, P =
0.023; Vx vs. Ux, χ2 = 6.017, P = 0.014.
Sperm retention without a Plug
Mating by males that did not produce plugs resulted in visible sperm masses in the
posterior portion of the female’s cloaca. Within 30 seconds post-copulation, a
considerable amount of sperm had leaked from the cloaca, leaving very few sperm in
the oviducts (Figure 3.2).
72
A
B
D
D
u
Figure 3.2: Two representative photos of slides wiped across the female’s cloaca within 30 u
r 20x: A: female mated with a Ux male that did not
r
seconds of the termination of copulation at
i
produce a plug and B: female mated with ai sham male that did produce a plug. It is clear that
n
n
there are no sperm in photo “B” and abundant
sperm in photo “A”.
g
g
Qualitative description of the plug and post-mating
cloacal
fluids
t
t
h
h
Histological sectioning showed increased sperm numbers toward the anterior end of the
e
e
s
s
plug (Figure 3.1). With the trichorme stain the plug matrix stains pink, indicating a high
p
p
r
concentration of sperm. The nuclei didrnot stain well with this method, but blue/green
i
i
n
indicates that the matrix is largely proteinaceous
(as suggested by Devine 1975). Then
g
g
e
H&E stain highlights the nuclei well and thus is a good indicator of sperm number e
m
m
e
(Figure 3.1).
e
r
r
g
g
Samples of fluid taken from the vaginal pouch of the cloaca after the plug was removed
e
e
n
n
had very few sperm. The openings of the paired oviducts were tightly closed in most
c
c
e
e
cases, and very few sperm were found in them via lavaging with ringer’s solution. We
o
o
conclude from these observations thatf most of the sperm are incorporated into the fplug
R
R
e dissolves within the female’s vaginal pouch.
e
matrix itself and are released as the plug
d
d
s
s
i
i
d
d
e
e
73
These are initial qualitative assessments. Quantitative analyses of sperm counts within
plugs are underway (Friesen & Mason in prep.).
In the “condom study”, we found no sperm in vials containing the mating plug, or in
flushings from the condoms that had been attached to unmated females. Mated
females with intact mating plugs yielded an average of 1,000 sperm, whereas those with
mating plugs removed yielded and an average of 73,000 sperm. Thus, sperm counts
differed significantly among treatments (ANOVA on ln-transformed data: F3,25 = 3.29, P <
0.04; Fisher’s post hoc PLSD tests show that the “plug removed” treatment resulted in
more sperm leakage than did any other treatment, with all P < 0.04).
Effect of multiple matings on plug deposition
In our arena mating trials, 20 males mated once, and of those 15 mated twice, and 10
mated a third time. Five of the thrice-mated males mated a fourth time, and one mated
a fifth time. There was no significant difference between sham and Vx treatments in
terms of (a) the total number of matings (Matings: t = 1.254, d.f. = 22, P = 0.223), (b)
total number of plugs produced (Plugs: t = 1.316, d.f. = 22, P = 0.202) or (c) plugs
produced per mating (plugs/mating U = 41.50, d.f. = 1, P = 0.516). Thus we excluded
treatment (Vx vs. Sham) from our analyses of the effect of multiple mating on plug
deposition. The number of matings per treatment differed primarily for the number of
third matings: 8 of 12 Vx males mated three times (= 67%) versus 2 of 12 Sham males
that mated three times (= 17%). If the analysis is limited to third matings only, this is a
74
significant difference, but it does not persist if a Bonferroni correction is applied to the
full model including all matings (χ2 = 4.286, d.f. = 1, P = 0.038; critical p-value with
Bonferroni correction = 0.008).
Nine males deposited one plug; six made 2 plugs; four made three plugs and one made
four plugs. The average number of plugs per male was 1.87, and the average number of
plugs per mating across all matings was 0.78 plugs/mating over six days of mating trials.
Plugs were produced in 19 of 20 first-matings (95%), and 12 of 15 second-matings (80%),
but only 4 of 10 third-matings (40%) and 1 of 5 fourth-matings (20%) and 0 of 1 fifthmatings (0%). We tested for the effect of male mate number on plug deposition with
several analyses. Given the binary response variable (plug or no plug), and the
imbalance in a repeated design, we first ran a multiple logistic regression on a limited
subset of the data using the 10 males that copulated 3 times or more and limited our
analysis to their first three matings (N = 30 observations among 10 males). In this
analysis, male mate number (Mate #) was a significant predictor of plug deposition but
male identity (ID) was not (full model: Likelihood ratio test statistic = 12.517, P = 0.002;
Mate #: Wald Stat. = 5.980, P = 0.014; ID: Wald Statistic = 0.0223, P = 0.881). Friedman
Repeated Measures Rank-sum test confirmed these results on the same data set (χ2 =
10.333, d.f. = 2, P = 0.006). Next, we included the data from all matings (excluding male
ID as a factor because of the previously mentioned analysis). We used maximum
likelihood (ML) based logistic regression in XLSTAT (Newton-Raphson algorithm) to
provide pairwise comparisons of rates of plug deposition with successive matings
75
[Model: -2 Log (Likelihood) χ2 = 17.88, P < 0.0001]. See Figure 3.3 for a comparison of
categories (male mate number on likelihood to produce a plug). All of our analyses show
that the likelihood of plug deposition drops off significantly after the second mating.
P = 0.004
P = 0.006
P = 0.199
1.0
P = 0.032
N = 20
N = 15
0.8
Proportion of plugs produced
P = 0.05
0.6
P = 0.447
N = 10
0.4
N=5
0.2
0.0
N=1
1st
2nd
3rd
4th
5th
Male Mating Number
Figure 3.3: Proportion plugs produced given the number of times a male mates. The likelihood of
producing a plug on successive matings decreases significantly after a males second mating.
Period between matings and plug deposition
We were also interested in whether the period between matings (Days) affected plug
deposition. However the relationship of mate number with days between matings is
76
complicated by possible autocollinearity (Spearman Rank Order Correlation (SROC): r =
0.704, P < 0.001). We used model selection methods, Akaike information criterion (AIC)
and Schwarz Bayesian information criterion (SBIC) to test which of these two variables
was more influential (in both methods lower scores indicate preferred models and are in
bold below); model: days between matings vs. null model of only the intercept AIC =
61.484 vs. 62.086; SBIC = 78.456 vs.73.401: model male mate number null model of only
the intercept AIC = 47.111 vs.62.086; SBIC = 64.083 vs.73.401). Accordingly, male mate
number is the preferred model using either criterion (days between matings was only
slightly preferred over the null using AIC, but not SBIC). The SBIC incorporates a larger
penalty for adding variables to the model and under this criterion days between matings
is not preferred over the null model that includes only the intercept term. Only one of
our multiply-mated males deposited a plug after not depositing one on his previous
mating (1st and 2nd matings). These first two matings by this one male occurred on the
same day, less than three hours apart. In this case the male’s first mating had a shorter
than average copulation duration (7:43 minutes vs. ̅ = 21:36), suggesting that
copulation duration might be an important predictor of plug deposition.
Copulation duration and plug deposition
To test the effect of copulation duration on plug deposition, we limited the analysis to
only those males that mated multiple times (i.e., males that mated 2-5 times; N = 46
matings among N = 16 males). This analysis confirms that copulation duration is a
significant predictor of plug deposition (mate #: Wald Stat. = 10.339, P = 0.001; Cop.
77
dur.: Wald Stat. = 4.665, P = 0.031; the full model Likelihood ratio test statistic = 23.673,
P < 0.001). Copulation duration was not significantly associated with male mate number
(SROC, r = -0.0552, P =0.714), so it appears that the effect of copulation duration is
independent of the mate number effect. Given the lack of an effect of multiple mating
on copulation duration in this data set, we collapsed all the matings in this data set to
test for a difference in copulation duration between copulations that produced a plug
versus those that did not. The average copulation duration of matings in which plugs
were produced ( ̅ = 23.16 minutes) was significantly longer than those that produced
no plugs ̅ = 18.38 minutes; t = 2.148, d.f. = 44, P = 0.037; see Figure 4). However, some
short matings produced a plug, suggesting that an experiment to determine the effect
of copulation duration on plug mass might be informative.
3000
2500
Copulation Duration (S)
2000
1500
Mean
Mean
1000
500
N = 31
N = 15
Plug
No Plug
0
Figure 3.4: Scatter plot of
copulation durations in
matings that did or did not
produce plugs (Plug or No
plug respectively). Bars are
standard errors.
78
No effect of male and female size on plug deposition
Neither female body size (SVL: Wald Stat. = 1.706, P = 0.192; Mass: Wald. Stat = 1.960, P
= 0.162) nor male body size (SVL: Wald Stat. = 0.428, P = 0.513; Mass: Wald. Stat =
0.104, P = 0.747) was significantly correlated with plug deposition.
Paternity analysis of a doubly-mated female
Paternity analysis of the single doubly-mated female revealed strong second-male
precedence in this case. The female gave birth to 12 offspring, nine of which (75%)
shared alleles that were present only in the second male while the other three (25%)
offspring shared alleles that were present only in the first male to mate. There was no
evidence of stored sperm; that is, all alleles matched the mother or one of the two
putative fathers. As this is only one litter, we can make no further general statistical
inferences to the population, but this anecdote is informative as it shows that deposited
plugs do not secure the first male’s paternity.
79
3.4
Discussion
The copulatory plug (CP) of red-sided garter snakes performs multiple functions. As
shown by previous studies, the CP of red-sided garter snakes can work as a chastity belt,
reducing the probability of re-mating by the female (for at least a few days: Shine et al.
2000a). However, our data suggest that mate-guarding may not be the only – or indeed,
primary – adaptive function of the copulatory plug in this species. Our experimental
studies show that the CP reduces leakage of sperm from the female reproductive tract:
sperm readily leaked from females after copulations in which no plug was deposited
(because we had ligated the male’s ureters) or after we removed recently-deposited
plugs. In contrast, sperm leakage was minimal from females after mating with males
that deposited CPs. Also, our study suggests another and perhaps equally significant
function: the CP may act as a spermatophore, gradually releasing sperm into the vaginal
pouch as the plug breaks down. Many other species of snakes show very prolonged
copulations (sometimes lasting for days: Olsson and Madsen 1998), which would fulfill
all of the same functions as we have identified for CPs in red-sided garter snakes.
However, the short mating season and availability of multiple receptive females imposes
a high “opportunity cost” to prolonged mating in this system, plausibly favoring a
reduction in duration of copulation (i.e., allowing a male to begin courting another
potential mate: Shine et al. 2000a). Our data suggest that male garter snakes can reap
the same fitness benefits as conferred by prolonged copulation, but without the
80
associated opportunity costs, by adopting the tactic of a brief mating followed by the
production of a large CP.
Our data confirm and extend previous analyses of this topic. For example, Shine et al.
(2000a) reported that the incidence of copulations without CP production increased
(albeit, non-significantly) if males engaged in multiple matings in quick succession. With
larger sample sizes and a more extended study, this latter trend attains statistical
significance in the current study. Additionally, we reveal an effect of copulation duration
on plug production, supporting the hypothesis that brief duration of copulation is
functionally linked to plug production. Copulatory plugs may be as important as sperm
numbers in the context of sperm competition in this system, because sperm that leaks
from the female’s reproductive tract cannot contribute to male fitness. Male red-sided
garter snakes clearly are limited in the number of plugs they can produce, hinting that
male snakes in this population may be under intense selection for rapid and repeated
plug production.
Ureter ligation
Although circumstantial evidence has pointed to the RSS as the source of plug material
for decades (e.g., Bishop 1959; Devine 1975), our study provides the first conclusive
support for that hypothesis in garter snakes; none of the ureter-ligated males produced
a plug. This technique, along with vasectomy, can be used to further elucidate the role
81
of CPs in assuring paternity and also, disentangling the effect of plug material and
seminal fluids on female mating behavior.
Testing the Sperm Leakage Hypothesis
The technique of ureter ligation allowed us to test a specific hypothesis about plug
function. Without the plug, sperm leaks from the female’s cloaca. Neither Devine (1975)
nor Shine et al. (2000a) completely disregarded the sperm leakage hypothesis, but they
instead focused on the mate guarding function. In the most comprehensive study on
plug function in garter snakes to date, Shine et al. (2000a) concluded that passive mate
guarding is the primary current function of the plug, and that seminal fluid, not the plug
per se, is responsible for reduced post-mating female attractivity. Their work
demonstrated that if the plug is removed, a female often will remate quite soon.
Nevertheless, the plug is not a totally effective chastity belt, because some females
contain multiple plugs (Shine et al. 2000a, and present study). Thus, the presence of a
CP does not necessarily render a female unattractive or unreceptive (Shine et al. 2000a;
Friesen pers. obs.). Cases of female red-sided garter snakes bearing multiple CPs calls
into question the ecological relevance of reduced post-mating female attractivity, when
large numbers of males are searching and competing for matings (e.g., Devine 1977;
Ross & Crews 1978; Shine et al. 2000). Furthermore, even in these cases the plug may
not protect the first male’s paternity. For example, male Iberian rock lizards produce a
copulatory plug, but this can be displaced by subsequent males (Moreira & Birkhead
82
2003); therefore, the plug does not ensure the first male’s paternity (Moreira, et al.
2007).
The copulatory plug turned spermatophore
We became interested in the question of whether the copulatory plug contained most
of a male’s sperm while developing a method for ejaculate collection, to calculate sperm
numbers per mating. Based on data from a single female, Devine (1975) suggested that
most of the sperm lie within the oviduct and vaginal pouch, and can be easily collected
and counted. Our data on plug composition (especially, the presence of abundant sperm
within it), and the paucity of sperm in the cloaca and oviduct, contrast with Devine’s
(1975) description. First, we found little sperm within the vaginal pouch. Second, in
partial agreement with Devine, we observed that there are more sperm in the anterior
sections than the most posterior sections of the plug; however, sperm were distributed
far more posteriorly than indicated by Devine’s schematic drawing (Figure 1). Devine
(1975) stated that “a far more homogeneous distribution of sperm in the plug would be
likely if the contents of the ductus deferens contributed significantly to the plug
material.” Our observations reveal exactly this pattern: a more homogeneous
distribution of sperm, consistent with material from the ductus deferens (i.e., sperm)
being embedded within the plug.
83
At least three factors may account for the discrepancy between our description and
Devine’s. First, the sphincter muscles surrounding orifices of the oviducts may control
the movement of sperm into the oviducts. These sphincters are usually closed (Friesen
pers. obs.). Semen is transferred first during copulation and the plug material second
(Shine et al. 2000a). Closure of these sphincters would prevent the sperm from entering
the oviducts. Sperm remaining in the vaginal pouch would then mix with the plug
material as it was subsequently deposited. This hypothesis could explain the distribution
of sperm within the plug that we observed. The animal examined by Devine may have
had a relaxed sphincter, allowing sperm could move up into the oviducts during
copulation.
A second possible explanation of the difference between our observations and Devine’s
is the elapsed time between plug deposition and collection. It is unclear how long the
plug had been in the female before Devine observed it, whereas we collected the plugs
immediately after deposition. If the plug had been in the female reproductive tract
longer, it may have already begun to dissolve and release sperm, which then traveled up
the oviducts. This would account for the “sperm-dense fluid” he found in the vaginal
pouch (that we have not found).
A third possible explanation is that the female may have mated previously and Devine
was observing the second plug. Thus the sperm he describes as “extending 10 cm up the
oviducts” might have been dissolved from a previous plug. Without controlling for the
84
number of times a female mated previously (as we were able to), we cannot rule out
this possibility. These are not mutually exclusive explanations; for example, a previously
deposited plug may contain substances that relax the oviductal sphincters.
Plug deposition
Ejaculates are costly for males to produce because they often contain millions of sperm,
accessory proteins and other substances that keep sperm viable and protect them from
the harsh environment within the female reproductive tract (Dewsbury 1982; Poiani
2006; Suarez 2006; Ramm, Parker & Stockley 2005). The copulatory plug of the redsided garter snake can represent more than 0.4% of a male’s mass (Shine et al. 2000a;
Friesen & Mason in prep), a substantial energetic investment (Friesen et al. in prep.).
Given sperm leakage after matings that do not involve a CP, males would gain little or
no paternity without plug deposition. Thus, prevention of sperm leakage may constitute
a major benefit to plug production.
Effect of multiple matings and copulation duration on plug deposition
A male snake’s prior mating history has a strong effect on plug deposition. Like Shine et
al. (2000a), we found no difference in plug deposition between first and second
matings. However, we found that the likelihood of plug deposition dropped off
significantly after the second mating; only 40% of third-matings and only 20% of fourthmatings produced a plug. The RSS remains hypertrophied throughout the spring
(Krohmer 2004), so presumably males have the capacity to produce more plug material.
Notably, males are aphagous during the spring breeding season and may be limited by
85
their energy stores (O’Donnell et al. 2004); they lose mass rapidly during this period
(Shine and Mason 2005). In addition to the energetic requirements of mate searching
and courtship, a male must maintain and allocate resources to plug production. We
were unable to distinguish the effect of the number of matings from the time between
matings on plug deposition (e.g., Oku & Kitsunezuka 2011), so we cannot assess
whether or how quickly males might replenish their stores of plug material. If they are
not able to replenish plug material when there are ample mating opportunities, we
might expect males to exhibit mate choice. In fact, males in this system do prefer to
court and follow trails left by larger, and therefore, more fecund females (LeMaster &
Mason 2002; Shine et al. 2006b; Shine 2012).
The observation that males are limited in the number of copulatory plugs they can
produce, suggests multiple further studies. For example, do males reduce courtship
activity when they are unable to produce a plug? We predict that males would curtail
courtship until plug material is replenished and then resume intense courtship.
However, males may not be able to assess their reservoir of plug material: some males
continued to court and mate without producing plugs (although our data hint at a
reduced mating frequency for such animals). As Vx males can mate without becoming
sperm-deleted, it would be interesting to use them to more explicitly test whether
sperm depletion was a cue to curtail courtship and mating effort. Further, sperm
numbers are important for success in sperm competition (Parker 1990), and in many
organisms copulation duration correlates directly with sperm numbers and can covary
86
with male size (e.g., Simmons & Parker 1992). Studies on correlates to variation in
sperm counts per ejaculate would clarify the targets of postcopulatory selection within
the red-sided garter snake system. Matings that produced plugs had longer copulation
durations than matings that did not. Copulation duration was not linked to a male’s
number of prior matings, but our subsequent studies (Friesen and Mason in prep.) have
found that copulations of less than 3 minutes rarely yield a plug.
King et al. (2009) interpreted copulation duration in garter snakes in terms of sexual
conflict. To the extent that the plug limits female remating, it is also a form of sexual
conflict (Stockley 1997; Arnqvist & Rowe 2005). Our study shows copulation duration
does have an effect on plug deposition, and that females would retain little sperm if the
plug was not deposited. Although the loss of sperm may seem like a disadvantage for
the female, she might thereby be able to void sperm from unwanted suitors (as do feral
fowl: Pizzari & Birkhead 2000). Given that female red-sided garter snakes have little precopulatory mate choice (Shine O’Connor & Mason 2000b; Shine, Langkilde & Mason
2003), voiding of “unwanted” sperm post-copulation might be the only mechanism by
which a female could influence paternity of her offspring. A copulatory plug would
prevent this form of cryptic female choice. In the context of sexual conflict, the plug may
function as a time release capsule to “out-wait” the oviductal sphincters which
otherwise would prevent sperm from advancing into the oviducts during copulation.
Thus, the sperm-leakage-reduction function of the CP might be a partial resolution of
sexual conflict in a system where female pre-copulatory choice is limited (Gowaty 1997;
87
Arnqvist & Rowe 2005). If there truly is conflict over plug deposition, then the
determinants of variation in copulation duration would be fertile ground for further
research.
Conclusion
Contrary to previous studies, our experiments suggest that the copulatory plug
produced by male red-sided garter snakes has a range of functions. It not only reduces
the probability of the female remating, but also reduces sperm leakage after copulation,
and functions as a spermatophore. All of these functions fit well with the idea that
specific features of the mating aggregations of red-sided garter snakes impose intense
selection against a long copulatory duration, and hence place a selective premium on a
male’s ability to transfer sperm quickly (Shine et al. 2000a). These multiple advantages
may explain why male snakes invest so many resources into such an ineffective “chastity
belt”. We do not know enough about the mating systems of related garter snake
species, or even of other populations within our study species, to disentangle the
evolutionary history of the copulatory plug. Plausibly, it initially evolved primarily to
fulfill one of the functions that we have identified, and then was co-opted for the
others. For example, passive mate guarding may be a spandrel (Gould & Lewontin 1979)
affixed to an adaptation to prevent sperm leakage. The presence of sperm in the plug
may be a proximate consequence of selection on female ability to control sperm
transfer, and ultimately reflect sexual conflict in these populations. Future work could
usefully explore the fitness benefits that accrue to males that invest additional
88
resources into larger and larger CPs; and also, to investigate the possibility of cryptic
female control over the use of sperm within a copulatory plug.
89
SPERM-DEPLETED MALES AND THE UNFORTUNATE FEMALES WHO
MATE WITH THEM
Christopher R. Friesen, Mattie K. Squire, Emily J. Uhrig, Deborah I. Lutterschmidt and Robert T.
Mason
Abstract
Female sexual promiscuity is prevalent element of mating systems. One consequence of
female sexual promiscuity is that male-male competition often continues postcopulation within the female’s reproductive tract. There are two central questions in the
study of postcopulatory sexual selection. 1) What factors determine male fertilization
success? 2) Why do females mate with multiple males? Explanations of female
promiscuity propose that indirect benefits (genetic; good genes, and bet hedging etc.),
and/or direct benefits (e.g., extra paternal care, obscuring paternity in social groups
etc.) to females outweigh the costs of female promiscuity (e.g., increased predation risk,
and exposure to STDs etc.). One hypothesized direct benefit is fertilization insurance;
females remate to ensure they have sperm to fertilize their eggs. We tested whether
female remating frequency was affected by mating with a sperm-depleted male. Our
results indicate that when female red-sided garter snakes (Thamnophis sirtalis
parietalis) mate with a vasectomized male, they are more likely to remate in seminatural arenas. Vasectomized males produced a copulatory plug, but did not deliver
sperm during mating. Two non-mutually exclusive hypotheses may account for our
results. 1) Females can sense sperm stores within the reproductive tract and use this
information to evaluate the quality of a recent mate and remate if the male was sperm
90
depleted. 2) The seminal fluid contains a substance(s) that lowers female receptivity to
subsequent matings. These hypotheses are intriguing, as the first suggests a form of
cryptic female choice, in which females remate after mating with a suboptimal male and
the second suggests sexual conflict in which males manipulate females to ensure their
own reproductive success.
91
4.1
Introduction
Females mating with multiple males whose sperm compete to fertilize ova (Parker 1970;
Simmons 2005) produce variation in male fertilization success (Dziuk 1996; Evans et al.
2003). Variation among males in ejaculate traits can account for some of this variation in
fertilization success (e.g., Birkhead et al. 1999; Boschetto et al. 2011). Differences
among males in the percentage of motile sperm, sperm morphology, velocity and
longevity all have been demonstrated to affect fertilization success under competitive
conditions (e.g., Birkhead et al. 1999; Miller and Pitnick 2002; Gage et al. 2004;
Casselman et al. 2006; Dziminski et al. 2009; Smith and Ryan 2010). However, according
to sperm competition theory, the best predictor of male fertilization success is the
number of sperm a male inseminates relative to his rivals (Parker 1990; Parker and
Pizzari 2010). Thus, when the risk of intense sperm competition is high, males should
invest heavily in their sperm and seminal fluid (ejaculates) (Parker 1990; Parker 1998;
Parker and Pizzari 2010). However, males may be limited in the number of sperm they
can produce because their ejaculates represent a substantial energetic expenditure
(Dewsbury 1982; Olsson et al. 1997). Further, sperm-depletion may be a factor in
species with compressed seasonal breeding periods, because the males have restricted
storage capacity and the time to replenish sperm stores is short (Wedell et al. 2002).
Consequently, the males of some species judiciously adjust their ejaculate to match the
quality of their mate and the conditions of sperm competition (e.g., Wedell and Cook
1999; Reinhold et al. 2002; Simmons et al. 2007; Lüpold et al. 2011; reviewed in Kelly
92
and Jennions 2011). Even so, a male that makes a large investment in one insemination
may be sperm-depleted for his next mating (Wedell et al. 2002; Smith et al. 2009).
If a female suffers from sperm-limitation (i.e. lack enough sperm to fertilize all of her
ova), then the possibility of mating with sperm-depleted males may increase the
tendency to remate (Wedell et al. 2002; Simmons 2005; Smith et al. 2009). Increased
female remating rates would lead to a further increase in the intensity of sperm
competition and greater investment in male ejaculates (Parker and Pizzari 2010). To
mitigate the costs of escalating investment in ejaculates, males may become more
selective of their mates and resort to mate-guarding behavior in order to safeguard
paternity (Dewsbury 1982; Wedell et al. 2002). Furthermore, males may differ in their
strategies of investment and allocation to ejaculates due to their condition, social
status, the female’s mating status (i.e., virgin or not) and the likelihood of their gaining
future matings (Parker 1990; 1998). Males that are less likely to gain matings have little
to lose by investing heavily in their first mating. Thus, patterns of paternity observed in
the wild are the result of both plastic behavioral responses and dynamic evolutionary
processes. Assessment of ejaculate traits, especially sperm numbers, may deliver
empirical data with which we can evaluate ultimate-level hypotheses that aim to explain
male and female reproductive behavior and previously established patterns of multiple
paternity (Simmons 2001).
93
Multiple paternity is widespread among reptiles, and snakes in particular, which
suggests that postcopulatory sexual selection is a pervasive factor in the group (Olsson
and Madsen 1998; Uller and Olsson 2008; Uller et al. 2010). Male red-sided garter
snakes (Thamnophis sirtalis parietalis), of Manitoba, Canada may be especially prone to
male sperm-depletion, female sperm-limitation, and divergent ejaculate-allocation
strategies among males. Males of this species display a dissociated reproductive
strategy in which sperm production occurs in August-September(Crews 1984) and
exhibit a compressed breeding period from late April through May (Gregory 1974). As
their testes regress during the breeding season (Krohmer et al. 1987), male T.s.
parietalis rely solely on their stored sperm. Nevertheless, males will mate many times if
given the opportunity (Friesen et al. submitted; Friesen et al. in prep.). As both
operational sex ratios and aggregation densities are high (Gregory 1974; Shine et al.
2000a; Shine et al. 2006), sexual selection on male traits is likely a prominent feature of
the mating system (Arnold and Duvall 1994; Shuster and Wade 2003; Klug et al. 2010).
In agreement with these theoretical predictions (Wedell et al. 2002), males of this
species produce a large gelatinous copulatory plug that functions to prevent sperm
leakage (Friesen et al. submitted) and reduce female remating rates, i.e., a passive
mate-guarding device (Shine et al. 2000b). The copulatory plug is produced by the renal
sexual segments (RSS) of the kidneys, and the fluids that form the plug have been
speculated to act as an antiaphrodisiac that reduces female receptivity (Ross and Crews
1977, 1978). In addition, males are somewhat selective of their mates, in that larger,
94
more fecund, females are courted preferentially and more intensely than smaller, less
fecund females (LeMaster and Mason 2002; Shine et al. 2003). Small males have an
equal chance to mate with a female as she emerges from brumation in the spring (Shine
et al. 2000a). However, in small to medium sized aggregations (2-20 males per female),
where second matings are most likely (Shine et al. 2001), larger males have a mating
advantage (Shine et al. 2000a). Thus, small males may use different ejaculate-allocation
strategies than larger males because they are in a disfavored role (Parker 1998).
To date, very few studies have assessed ejaculate traits in snakes (Schutle-Hostedde and
Montgomerie 2006; Tourmente et al. 2006; Fahrig et al. 2007; Mattson et al.
2007;Tourmente et al. 2007; Tourmente et al. 2009; Tourmente et al. 2011). Most of
these studies only addressed sperm morphology, and all of these studies collected
semen samples by hand-manipulation or from museum specimens. Here we examine
ejaculate traits [mobility (e.g., Froman et al. 1999), and total sperm numbers] from
natural inseminations in a well-studied model system, the red-sided garter snake of
Manitoba, Canada. We tested whether sperm-depletion occurs by comparing sperm
counts and ejaculate quality of first and second matings. We also assessed the effect of
male size on sperm numbers and ejaculate quality, as smaller males of many species
increase investment in ejaculates to compensate for lower chances of mating (e.g.,
Simmons and Parker 1992; Bissoondath and Wikilund 1996). We evaluated the effect of
female size on sperm numbers and ejaculate quality, to test the hypothesis that males
prudently adjust their ejaculate in response to the quality of their mates (Wedell et
95
al.2002). Finally, we tested whether female remating-rates increase when they are
mated with a male that has been vasectomized and thus, whose ejaculate is completely
devoid of sperm.
4.2
Methods
Model system
Red-sided garter snakes are small (adult males average 45 cm in snout-vent length
[Msvl], and females 68 cm Fsvl;), non-venomous natricine colubrids. Larger females
produce larger litters (Larsen and Gregory 1993). Our study population is located near
Inwood, Manitoba, Canada (50° 31.58’N 97°29.71’W). This population contains
approximately 35, 000 individuals (Shine et al. 2006). Males emerge from underground
winter brumation sites in late April and form large, dense aggregations around the
emergence sites. As females emerge they are met with the amorous attention of 3-62
males (Shine et al. 2006). One hundred actively courting males and 50 newly emerged
seasonal virgin females were collected by hand the day before mating trials which began
on 19 May, 2009. The animals were brought to Chatfield Research Station, 16 km north
of the collection site, and housed in seminatural nylon arenas and provided water ad
libitum. Males and females were housed separately until the mating trials began. In
addition, thirty males were collected from Snake Island population [51° 38.53’N 99°
49.42’W (20 May 2012)] and mating trials were conducted two days later for sperm
counts at Chatfield Research Station, MB. Finally, 48 males and 48 females were
96
collected from the Inwood population in May of 2012 for the female remating rate
experiment.
Ejaculate collection
Small circular arenas (45 cm dia. x 75 cm tall) were setup indoors at the Chatfield
Research Station with each placed under a 250 watt heat-lamp 1m above the animals.
We measured external body temperatures with a laser thermometer and adjusted the
output of each lamp individually using dimmer switches to maintain optimal body
temperature at 29-30°C (Kitchell 1969; Hawley and Aleksiuk 1975). Twenty males were
randomly assigned to each of the arenas and were allowed to court and mate with
unmated seasonal virgins that were collected the same day as the male. A sex ratio of
twenty males to one female is common in and around the dens (Joy and Crews 1985;
Shine et a;. 2004). Courtship was observed closely to facilitate the timing of copulation
duration (±10s) when mating commenced. After copulation was initiated and had lasted
one minute, the pair was removed to a separate, empty, circular arena so that they
would copulate without interference from the other males; this separation also allowed
easy observation of the termination of copulations. Thirty of these males mated on the
first day (first matings) and these 30 males were allowed to court and mate with
another set of unmated seasonal virgin females on the next day. Of the 30 males that
mated on the first day, 15 mated a second time the next day (second matings) and three
of those males mated a third time.
97
Less than 30s after copulation terminated, each female was inspected for a copulatory
plug. Each plug was removed by gently running a blunt probe around the plug to
separate it from the walls of the vaginal pouch. Once removed the plug was placed in
1.5 ml microcentrifuge tube in 1ml of Modified Ham’s F-10 medium and 10 µg/ml of the
antibiotic Gentamicin Sulfate (Cat # 99175, Irvine Scientific; e.g., Mattson et al. 2007).
The females’ vaginal pouch was lavaged with the same Ham’s F-10 medium using a
20ga. intubation needle affixed to a 1ml syringe. The fluid from the vaginal-wash would
contain any sperm not embedded within the plug and was subsequently added to the
1.5ml tube with the plug. The tubes were placed in a refrigerator for two days and were
gently agitated three times daily to aid the liberation of sperm embedded within the
plug. The dissolution of the plug was evidenced by a dense “cloud” of sperm above the
plug. When most of the sperm were liberated, a small piece of the posterior portion of
the plug remained. This portion contains very little sperm (Friesen et al. submitted). In
this way, we collected 45 ejaculates for sperm counts and the mobility assay from the
Inwood population 2009, and 14 ejaculates for sperm counts from the Snake Island
population in 2012. The Snake Island males were mated with newly emerged Inwood
females (21 May, 2012).
Mobility assay
The mobility assay measured the ability of a population of sperm cells to swim against
resistance through dense medium, in this case 3% (wt/vol) Accudenz® in Modified
Ham’s F-10 medium. The number of sperm that have adequate velocity to penetrate the
98
medium is proportional to the absorbance. Absorbance was measured using a portable
spectrophotometer (ARS 596A Sperm Mobility Analyzer) at 550 nm, which correlates
with the number of sperm that penetrate the medium. On the third day after the mating
trials, we conducted a mobility assay modified from Froman and McLean (1996; Froman
2006). The sperm samples were kept in a cooler with ice packs until one hour before the
mobility assay was conducted, and then placed in a 28-30°C water bath for an hour
before absorbance readings were recorded. Volumes (1.5ml) of 3% (wt/vol) Accudenz®
were pipetted into each of two standard polystyrene cuvettes, each cuvette was
covered with 1 cm2 piece of Parafilm®, and placed in the 28-30°C water bath in
preparation for the assay. Bubbles were gently tapped from the Accudenz® just prior to
blanking the cuvette in the spectrophotometer. The ejaculate sample in the 1.5 ml tube
was gently inverted three times to mix the sperm, and then set aside for 30s to allow
undissolved plug debris to settle. A 150 µl sample of the “ejaculate” was overlaid on the
surface of the Accudenz® solution, and the cuvette was returned to the 28-30°C water
bath. After a 5 min interval, the cuvette was transferred to the spectrophotometer and
absorbance at 550 nm recorded and again at a 10 min interval (T2). Finally, we made
one final absorbance reading after vigorously mixing (“Mix”) the contents of the cuvette
to control for variation in sperm numbers within and among samples. This procedure
was repeated with the same sample for a duplicate reading.
99
Preliminary analysis of sperm mobility
The average within-pair coefficient of variation (SD/ ̅ ) for all pairs of absorbance
readings (N = 35 pairs) was 0.061. Absorbance readings of mixed samples (Mix) and
sperm counts for each replicate cuvette were highly correlated (R = .921, Adj. R 2 = 0.846,
P < 0.001).The plugs were allowed to dissolve for the same period (± 1h), i.e., the
interval between mating and when the assay was run, and not correlated with the start
time of the assay (R2 = 0.018, P = 0.540). However, start time was positively correlated
with absorbance readings at 5 min. (T1: R2 = 0.236, P = 0.019) but not 10 min. (T2: R2 =
0.146, P = 0.072) or those of the samples after mixing (Mix: R2 = 0.006, P = 0.732). The
effect of start time became pronounced after controlling for sperm numbers, i.e., T2
absorbance/Mix absorbance (T2/Mix: R2 = 0.414, P = 0.001). It is possible the cooler was
slightly warmer than the refrigerator which allowed samples run later in the day to
warm slightly before the assays were performed. Therefore, the effect of start time of
the assay was statistically removed by using the standardized residuals from a
regression analysis of T2/Mix as a function of start time for the remainder of the
analyses.
Sperm counts
After absorbance was measured, the contents of each cuvette and the original tube
containing the plug were collected in separate 1.5 ml tubes, centrifuged for 10 minutes
at 600 rpm. Sperm from each cuvette and the tube containing the plug were preserved
for counts in 3% final PFA in PBS. Sperm counts were made in triplicate for each cuvette
100
as well as the tube containing the sperm and the plug using a Petroff-Hausser Counter
(cat. # 3900, Hausser Scientific). The total number of sperm in each ejaculate was
estimated as the sum of the average counts for each replicate cuvette and the tube
containing the plug. The plug mass was not recorded immediately after copulation
because we felt it was important to minimize handling of the ejaculate for mobility
analysis. However, once the sperm were liberated (dissolution), the mass of the
relatively sperm-free posterior portion of the plug was recorded.
Female remating rates
Vasectomies
During May 2012, 48 vigorously courting males were collected from the study
population. These males were randomly assigned to one of two treatment groups:
vasectomy (N = 24) and intact controls. In squamate reptiles (e.g., lizards, snakes) each
testis has a separate duct that conveys the sperm to one of two hemipenes (Fox 1977).
Likewise, each kidney has a separate duct that carries secretions produced by the renal
sexual segment (RSS), which produces the copulatory plug (Friesen et al. in prep.). To
simulate mating with a sperm-depleted male, we used females that received the RSS
sections and a copulatory plug but not sperm (i.e., simulated mating with a spermdepleted male). Others have attributed antiaphrodisiac qualities to the RSS secretions
(Ross and Crews 1977; 1978). To test this hypothesis we staged mating with
vasectomized (Vx) males; these males were unable to transfer sperm, but could still
deposit a copulatory plug since their RSS ducts (ureters) remained intact. Vx males
101
produce copulatory plugs at the same rate as either sham-surgery or intact control
males (Friesen et al. submitted, and Friesen unpublished data). We have shown
previously that vasectomized (Vx) males show no deficits: they court vigorously and
mate within 24 h of surgery (Friesen et al. submitted). Thus, we performed 24
vasectomies (Vx) and used unmanipulated control males for mating trials.
Before surgery, 15 mg/kg of anesthetic (0.003 ml of 0.5% methohexital sodium/g body
mass) was administered subcutaneously at the juncture between the dorsal and ventral
scales at a distance of approximately 20 cm from the head of the snake (Preston et al.
2010). Upon anesthetization, an incision was made in the skin between the first and
second lateral scale rows directly dorsal to the 5th ventral scale (counting from the vent
scale). The surgical window traversed no more than four scales and the incision was
made from posterior to anterior with sterilized corneoscleral scissors. The peritoneal
membrane was then cut using the same scissors and made wider than the skin window
to allow for easier access to the ductus deferens. Both left and right ductus deferens
were isolated from the viscera and ligated using sterile suture (4-0 G silk). The surgical
window was then sutured shut using sterile suture (4-0 G silk). Subsequently, the
incision site was first rinsed with Ringer’s solution (reptile physiological saline), then
swabbed with 100% ethanol to prevent infection. The snakes were then placed in an
aquarium with a heating pad (30°C) and monitored every 10 minutes for the first hour
after surgery until righting reflex returned. Males of this species regain their drive to
102
court very soon after surgery and show no deficit in this ability (e.g., Nelson et al. 1987).
All animals survived surgery and engaged in courtship within 24 hours after recovery.
Vx mating trials
Mating trials were conducted in the same manner as for ejaculate collection described
above. Forty-three newly emerged seasonal virgin females mated with either an intact
control or Vx male (83% of both the control and Vx males mated over two days, 5/4-5/5
2012). Three of these matings to Vx males did not produce plugs; this was most likely
due to these being second matings for each male. The three females that did not receive
plugs during mating were removed from the experiment. There were 40 females mated
and with copulatory plugs available for the second mating trials (Vx, N = 20; Ctrl, N = 20).
Each plug was marked with green food coloring (FD&C Green No. 3) to identify it as the
original plug in subsequent unobserved matings (the food coloring stains the plug until it
fully dissolves with the female’s cloaca, and does not affect viable sperm, Johnson and
Welch 1999; Friesen unpublished data). The females were then placed in a natural
outdoor enclosure (3.65m diameter x 1m, 11.5m2). The floor of the enclosure was grass
and brush like that of the aspen grove that surrounds the Inwood den site. Females and
males form small mating balls after females mate in the den in this type of substrate
(Shine et al. 2001). Water was provided ad libitum in a 0.3m diameter hole dug in the
floor of the enclosure that was lined with a clean sheet of plastic. One hundred
vigorously courting males from the nearby den (< 50m away), were placed in the arena
with the mated females. Females were collected and checked for new plugs every day
103
for six days (a freshly deposited plug will dissolve within the female’s cloaca in two days,
Shine et al. 2000b), thus we were sure to note if females mated again as evidenced by a
fresh (not green colored) copulatory plug. The Vx males cannot be released due to
permit requirements, so they were sacrificed to assess parasite loads for another
experiment. During these dissections, it became apparent that four of the Vx males had
unilateral vasectomies. It was not clear whether these males delivered sperm or not, so
the females that mated with them were removed from the experiment (three of the
females had remated and one did not).
104
4.3
Results
Sperm numbers
Five plugs were excluded from the analysis due to excessive clumping due to incomplete
deliquescence of the plug, which precluded repeatable sperm counts. Sperm counts
were estimated from 45 plugs produced by males from the main study population and
ranged from 1.19x107 to 7.01x108 [ ̅ = 7.98x107 (SEM = 1.89 x107)]. Sperm counts
conducted in 2012 from plugs produced by 14 males from the Snake Island population
were not significantly different from all 45 collected from Inwood males in 2009 (t df=41 =
1.673, P = 0.102). Sperm counts were not affected by male size (R2 = 0.004, P = 0.673) or
copulation duration (R2 = 0.000, P = 0.994) Figure 4.1a. The number of sperm
inseminated and copulation duration were not correlated (R2 = 0.027; P = 0.451) Figure
4.1b, and copulation duration did not depend on male size (Msvl, R2= 0.015, P = 0.523).
Males did not allocate more sperm to larger females (Fsvl, R2 = 0.047, P = 0.322).
However, copulation duration was positively correlated with female mass (Fmass, R2 =
0.187, P = 0.024). There was weak evidence of a positive, but non-significant, correlation
between the mass of the plug after liberation of the sperm (hereafter, plug mass) and
female size (Fsvl, R2 = 0.158, P = 0.060), and plug mass was significantly correlated with
copulation duration (R2 = 0.221, P = 0.018). Plug mass was not related to the number of
sperm inseminated (R2 = 0.011, P = 0.611).
105
21
21
b
20
20
19
19
Ln(sperm count)
Ln(sperm counts)
a
18
17
16
18
17
16
15
15
14
14
13
13
36
38
40
42
44
46
48
50
52
54
56
200
400
600
800
1000
1200
1400
1600
Copulation Duration (s)
Msvl
Figure 4.1: Regressions plots of the relationship between sperm numbers from natural ejaculates
2
and a) male size (Msvl; R = 0.004; P = 0.673) b) The relationship between sperm inseminated and
2
copulation duration (R = 0.027; P = 0.451).
21
Sperm numbers decreased significantly
20
from a first mating to the second, with first
than second matings (N = 12, Repeated
measures ANOVA, F1, 16 = 13.011, P = 0.004),
Figure 4.2. Sperm numbers from three
Ln(Sperm Counts)
matings producing 5.8 times more sperm
19
18
17
16
15
14
13
plugs produced from third matings were
4.75x106, 8.61x107 and 3.66x 108.
1
Male Mate #
2
Figure 4.2: Before and after line plot of the
decrease in sperm numbers from the first to
second matings (Repeated measures
ANOVA, F1, 16 = 13.011, P = 0.004).
106
Limiting the analysis to the male’s first matings only, male size (snout to vent length: svl)
and body condition (BCI: residuals of Mmass | Msvl generated by regression analysis)
did not predict sperm numbers within the Inwood (Msvl, R2 = 0.105, P = 0.131; BCI, R2 =
0.025, P = 0.475) or Snake Island populations (Msvl, R2 = 0.048, P = 0.538; BCI, R2 =
0.000, P = 0.950). However, Inwood males inseminate significantly more sperm on their
first mating than do Snake Island males (Mann-Whitney U = 83.00, P = 0.007).
Sperm mobility
The initial analysis was confined to the male’s first matings. Seven individuals were
removed from the analysis due to unusually high variance in absorbance readings that
were attributable to debris from plug material. Time of day when the assay was
conducted had a significant effect on mobility (T2/Mix) for first matings (R2 = 0.341, P <
0.001), thus standardized residuals from this regression analysis were used for the
remainder of the analyses. Among-male variation in mobility was significant (KruskalWallis ANOVA on ranks; H = 37.265, df = 22, P = 0.022), Figure 4.3. AIC model selection
of the (first matings only) revealed a significant negative relationship between sperm
quality and male size (Msvl: R2 = 0.122; P = 0.042), Figure 4.4. Mobility was not
correlated with total sperm inseminated in (R2 = 0.035, P = 0.392) or BCI (R2 = 0.039, P =
0.324).
3
3
2
2
Standardized Sperm Mobility
Standardized Sperm Mobility
107
1
0
-1
-2
1
0
-1
-2
Males
Figure 4.3: Intrapopulational variation in
ejaculate quality as measured by standardized
mobility score of first matings.
36
38
40
42
44
46
48
50
52
54
56
Msvl (cm)
Figure 4.4: Effect of male size on sperm mobility
of first matings.
108
Mobility of first and second matings by 15 males was analyzed to assess sperm quality
after successive matings (male mate #). Mobility improved significantly from first to
second matings [GLM mixed model (REML)], with mate number as a fixed effect, male
indenity as a random effect and start-time of assay as a covariate; F1, 29 = 10.530, P =
0.003), Figure 4.5. Female size was not significantly different between first and second
matings (Fsvl, tdf=14 = 0.981, P = 0.343) and the difference in female size (Fsvl) from the
first mating to the second was not correlated with the change in a male’s sperm mobility
(R2 = 0.022, P = 0.595). Male size was not significantly correlated with mobility in the
second mating (R2 = 0.000, P = 0.856).
3
Standardized Sperm Mobility
3
a
2
2
1
1
0
0
-1
-1
-2
b
-2
36
38
40
42
44
46
48
50
52
54
56
1 Male Mate # 2
Msvl (cm)
2
Figure 4.5: a) Relationship between male size (Msvl), first matings (open triangles and dotted line; R =
2
0.110, P = 0.123), second matings (closed circles and solid line; R = 0.000, P = 0.856) and sperm
mobility. The slopes are not significantly different (ANCOVA, interaction of Msvl x Mate #, P = 0.198). b)
Before and after line plot of the increase in mobility from the first to the second matings (GLM mixed
model (REML), with male mate # as a fixed effect, male ID as a random effect and start-time of assay as
a covariate; F1, 29 = 10.530, P = 0.003).
109
Vx mating trials
Of the 20 females that mated with Vx males, 14 (70%) remated. However, we had to
remove four of these females from the analysis because the males they mated with
were unilaterally vasectomized (i.e., only on duct was ligated); thus 11 of 16 (68.8%)
females mated to Vx males remated. Of the 20 females that mated with intact control
males, 4 of 20 (20%) remated. Significantly more Vx mated females remated than
females mated with controls (Yates continuity corrected (YCC) χ2df 1 = 6.801, P = 0.009).
If we are extremely conservative and add the four unilateral Vx mated females to the
control group (11/16 rematings Vx; 7/17 rematings controls) the difference in remating
rate is still significant (YCC χ2df 1 = 4.583, P = 0.032).
110
4.4
Discussion
This study rocks because…. Our study is the first investigation of ejaculate traits from
natural inseminations in a snake. All male T.s.parietalis invest heavily in sperm and may
tailor the size of the plug with increased copulation durations in an effort to defend
their paternity with larger females. Small males invest more in sperm relative to their
body size despite having smaller absolute testes mass. In the largest mating
aggregations where precopulatory female choice seems limited, copulation may be a
crucial period when females may assess their mate. Indeed, variation among males in
sperm numbers and copulation duration may be the result of female x male interactions
during copulation. Females then, may want to trade-up on a second mating. Copulations
are longer than required for insemination may only increase plug mass. As the plug
reduces the female’s chances of remating, copulation duration may be a source of
sexual conflict.
We investigated within-population variation in sperm traits within the framework of
sperm competition theory and female promiscuity. The ejaculate of T.s. parietalis
consists of tens to hundreds of million sperm that are highly concentrated within a
copulatory plug, which prevents the loss of this large investment (Friesen et al.
submitted). We found strong evidence that males cannot maintain the same
expenditure of sperm from one mating to the next. There was over a five-fold decrease
in sperm numbers from first to second matings. The mate-order effect of declining
sperm numbers might translate to a decline in paternity, however, in a previous
111
experiment; we found no evidence for a decline in paternity across successive male
matings (Friesen et al. in prep.). Nevertheless, when faced with sperm competition a
five-fold disadvantage could limit male fertilization success. Without knowledge of
sperm attrition rates within the female reproductive tract, we cannot know when a
male should be considered sperm-depleted. Twenty percent of females remate in
arenas (Shine et al. 2000a), and if they are remating to receive sperm, then this
percentage may represent the fraction of males that are sperm-depleted.
Within the context of sperm-depletion, it is interesting that ejaculate quality, as
measured by sperm mobility, improved on second matings. An increase in sperm
mobility corrected for the number of sperm (i.e. T2/Mix) represents a larger proportion
of sperm that moves against resistance (Froman and Feltmann 2000), which has been
shown to predict sperm competiveness (Birkhead et al. 1999; Pizzari et al. 2008). It is
conceivable that the improvement in quality makes up for a decline of sperm numbers
from the first mating to the second in a competitive context. Thus, an increase in the
quality of the ejaculate even with a decrease in sperm numbers may explain why
paternity was unaffected our previous result of mate number (Friesen et al. in prep.). A
mechanistic explanation for the increase in sperm quality seems straightforward. In T.s.
parietalis, spermatogenesis occurs in August through September (Krohmer et al. 1987)
and as sperm are produced they first fill the most caudal portion of the ductus deferens
(Friesen per. obs.). These sperm would also be the first to be inseminated. Sperm
function declines with sperm age (Pizzari et al. 2008), and earlier inseminations probably
112
contain older sperm. In first matings, smaller males had higher mobility than larger
males, but this effect disappeared in second matings. Larger males may have made
more sperm earlier during spermatogenesis, and thus have more “old” sperm relative to
new sperm in their first ejaculate. Given the lack of the same trend in mobility for
second matings this seems the most plausible explanation.
Larger male T.s.parietalis have larger testes (Friesen et al. unpublished data, Appendix),
so larger males might inseminate more sperm than smaller males. However, although
there was huge variation among males, inseminations contained the same absolute
number of sperm regardless of male size. In at least one species of snake species,
Nerodia sipedon, sperm concentrations are higher in smaller males (Schulte-Hostedde
and Montgomerie 2006); thus, smaller males may be compensating for fewer mating
opportunities. In some species the rate of insemination is higher in larger males and
thus larger males are able to inseminate an equal number of sperm as smaller males in a
shorter period (e.g., crickets, Simmons and Parker 1992; lizards, Olsson 2001). However,
smaller males did not copulate longer than larger males. Furthermore, sperm numbers
decreased equally from first to second matings regardless of male size, thus larger males
did not save sperm for second matings. Relative to their body size, smaller males do
seem to invest more in their ejaculates than larger males.
Males did not adjust the quantity or the quality of their sperm in accordance to female
size or mass. However, copulation duration was correlated with both the relatively
113
sperm-free portion of the plug and female mass. There was a weak insignificant
relationship between plug mass and female mass. Shine et al. (2000b) found a similar
trend between whole plug mass and female size, but no relationship with copulation
duration. Most of the sperm inseminated is contained within the plug but is mostly
distributed in the anterior to medial sections (Friesen et al. submitted). Given the lack of
relationship between sperm numbers and copulation duration, it may be that the
variation in copulation duration is explained by allocating more plug material to larger
females. To the extent that plugs act as passive mate-guarding devices, males may be
allocating more mate-guarding resources to larger females instead of more sperm.
Males may inseminate as much sperm as they are able during their first mating
regardless of female attributes.
Among male variation in ejaculate may explain the variation in paternity patterns that
we have uncovered in a previous study (Friesen et al. in prep.). One source of variation
is whether males mated before collection. All of the males we collected had seemed to
be newly emerged, had good body condition and were vigorous courters. However, one
could not know if a newly emerged male had mated the previous fall, which may
account for variation in both sperm quantity and quality. Schutle-Hostedde and
Montgomerie (2006) also failed to find significant effect of body condition on sperm
concentration in water snakes. Parasite loads and intrinsic resistance and tolerance of
parasites may play a role in determining sperm quality and quantity, and deserve
immediate investigation.
114
Females that emerge in the spring cannot be any more certain of the mated status of
her mate than we were. Consequently, they risk mating with sperm-depleted males.
Females are much more likely to remate if they do not receive seminal fluid and sperm
from the ductus deferens regardless of RSS secreted plug deposition. Two mechanistic
hypotheses may account for the increased remating rates: 1) the seminal fluid from the
ductus deferens may contain an antiaphrodisiac as Ross and Crews (1977; 1978)
suggested were contained within the RSS secretions; 2) females may have some sort of
proprioception of the sperm themselves that they use as a cue to seek out another
mating. These two hypotheses are not mutually exclusive. To adequately distinguish
between them, one needs to inseminate females with sperm washed of all other fluids
within the ductus deferens, and then mate her with a Vx male to produce a plug. In
addition, the fluid of the ductus deferens, sans sperm, would need to be inseminated
into females that are also mated with Vx males.
Our experiment using Vx-mated females to examine remating rates establishes the
foundation for several lines of inquiry such as if “females ever sperm-limited?” Females
of some species are sperm-limited when they mate with sperm-depleted males
(reviewed in Wedell et al.2002). It is a fair critique of our experimental design to suggest
that using vasectomized males is an exaggerated degree of sperm-depletion. We do not
know if females are ever sperm-limited in the actual mating system of our population. If
some males are systematically more successful, then those males may be the males that
become sperm depleted and the females they mate with may be sperm-limited. Male
115
mating success in the field has not been determined in this population, so no estimate
of the probability of mating success is available to evaluate this part of male mating
strategy. However, in arena trials, males will mate multiply if given the opportunity and
some males are more successful than others. A follow-up experiment would be to mate
males multiple times to collect their ejaculates and establish sperm-depletion over
many successive matings. In a parallel experiment, the effect of female sperm-limitation
could be addressed by testing for an effect of male mate number on probability to give
birth to viable offspring. This experiment is complicated by long-term sperm storage so
paternity analysis would be needed to confirm the focal male fathered offspring.
Finally, we compared sperm numbers in two populations. Snake Island males
inseminated fewer sperm than Inwood males on first matings. Three important caveats
are: 1) the ejaculates were collected in different years; 2) the mating history of the
Snake Island males is completely unknown; and 3) they mated with Inwood females.
Inwood and Snake Island populations differ in population density, and show sperm
precedence patterns indicative of superior sperm competitive ability (Friesen et al. in
prep.). The disparity in sperm numbers may account for this previously described
pattern in paternity. Corroborative evidence for the difference in sperm numbers lies in
the fact that Snake Island males have smaller relative testes mass than Inwood males
(Friesen et al. unpublished data, Appendix). Differences in the mating system, i.e.,
higher mating aggregation density of the Inwood population, and thus potential levels of
sperm competition and sexual conflict may explain increased testes mass and sperm
116
numbers in the Inwood population. However, we cannot rule out the possibility that
other ecological factors may be at the root of these differences.
117
Appendix to chapter 4
0.0
-0.5
Ln(testes mass)
-1.0
-1.5
-2.0
-2.5
-3.0
-3.5
3.5
3.6
3.7
3.8
3.9
4.0
4.1
4.2
Ln(Msvl)
Figure 4.6: Testes mass vs. Male size of 23 Snake Island (closed circles and solid regression line) and 34
Inwood males (open triangles and dashed regression line) collected for characterization of parasite
loads summer 2012 (ANCOVA: F 3, 56 = 21.13, P < 0.0001, Inw. Vs. Snk Bonferroni corrected P = 0.008,
Pop x Msvl interaction P = 0.548).
0.6
0.4
20
18
0.2
Ln(Sperm counts)
Residual Testes mass (g) | Msvl (cm)
22
0.0
16
14
-0.2
12
-0.4
10
SI
IN
Populations
Figure 4.7: Residual testes mass of 23 Snake
Island and 34 Inwood males (Mann-Whitney U =
244.00, P = 0.017) summer 2012. The boxes
enclose 50% of the data, the whiskers are 1.5
interquartile range, and the solid line is the
median.
Snk
Population
Inw
Figure 4.8: Sperm counts from 14 plugs collected from
Snake Island males that mated with Inwood females
late spring 2012 and 25 Inwood males spring 2009
(Mann-Whitney U = 83.00, P = 0.007). The boxes enclose
50% of the data, the whiskers are 1.5 interquartile
range, and the solid line is the median.
118
FEMALE CONTROL OVER COPULATION DURATION AS EVIDENCED
BY MALE GENITAL MANIPULATION AND LOCAL ANESTHETIZATION
OF THE FEMALE VAGINAL POUCH IN GARTER SNAKES
Christopher R. Friesen, Emily J. Uhrig, Mattie K. Squire, Robert T. Mason and Patricia
L.R. Brennan
Abstract
Sexual conflict is when the evolutionary interests of females and males are divergent.
Sex-differences in optimal copulation duration can be a source of conflict. Males may
evolve mechanisms to their reproductive success which prevent females from remating
to ensure, while females may otherwise benefit from mating again with a different
male. Increased copulation duration may be advantageous for males as it delays female
remating. Males of many species actively guard females to prevent them from remating,
and in some cases males produce copulatory plugs to prevent remating. This conflict
may be especially onerous to a female if precopulatory choice is limited at the time of
her first mating. Male red-sided garter snakes (Thamnophis sirtalis parietalis) produce a
gelatinous copulatory plug during mating that occludes the opening of the female
reproductive tract for approximately two days. The size of the plug is influenced by the
copulation duration. We experimentally tested the contribution of male and female
control over copulation duration. We ablated the largest basal spine on the male’s
hemipene and found a reduction in copulation duration and an increase in the variation
of plug mass. Further we anesthetized the female’s cloaca and found copulation
duration increased in this treatment group as well. This suggests that males benefit
from increased copulation duration while females actively try to reduce copulation
119
duration. Therefore, sexual conflict is manifest in divergent copulation duration optima
for males and females.
120
5.1
Introduction
Female polyandry is widespread among animal taxa, which suggests that females may
receive benefits from mating with multiple males (Jennions and Petrie 2000; Hosken and
Stockley 2003; Simmons 2005; but see Halliday and Arnold 1987). In species which
exhibit female sexual promiscuity, after mating with a female, males may attempt to
delay or prevent her from mating again to ensure his paternity in the face of sperm
competition (Parker 1970b; Smith 1984; Stockley 1997; Birkhead and Møller 1998). This
may be accomplished by physically guarding females, occluding the opening of the
female reproductive tract with a plug or intromittent organ, prolonged or repeated
copulations, or even removing the female from breeding areas (e.g., Parker 1970a;
Shine et al. 2000a; reviewed in Andersson 1994; Birkhead and Møller 1998; Simmons
2001).
Sexual conflict predicts that adaptions and counter-adaptations by the sexes will arise
through a dynamic evolutionary arms race in which one sex limits the evolutionary
interests of the other sex (Parker 1979, 1984; Arnqvist and Rowe 2005). Male control
over female remating is a form of sexual conflict if it limits females from attaining their
fitness optima (Parker 1979, 1984; Stockley 1997; Arnqvist and Rowe 2005). Females are
not evolutionarily passive and may evolve responses to sexual conflict ranging from
behavioral adaptations of cryptic female choice such as sperm ejection in fowl (Pizzari
and Birkhead 2000) to morphological adaptations such as the multiple blind-ended
invaginations within the female reproductive tracts of waterfowl (e.g., Brennan et al.
121
2007; Brennan et al. 2010). The males of these waterfowl species that force copulations
have highly modified genitalia that explosively intromit into the female cloaca (Brennan
et al. 2010).
Copulatory plugs, then, may be a form of conflict if they evolved to prevent females
from remating (Voss 1979; Shine et al. 2000a). Following the logic of an evolutionary
arms-race scenario, if the deposition of copulatory plugs is a form of conflict, then we
might expect female counter-adaptations, such as resistance to plug deposition. Males
might then counter female resistance with further adaptation. This scenario suggests
that copulation duration may be a source of conflict as it determines the size of the
portion of the plug that functions to reduce female remating (Friesen et al. in prep.) and
that male and female garter snakes may be locked in an antagonistic arms race over
plug deposition. For instance, the male genitalia of species that deposit copulatory plugs
may exhibit morphological adaptations that facilitate plug deposition especially if
females resist. Indeed, male intromittent organ morphology has been correlated with
male fertilization success, forced copulations, and female harm during copulation
(House and Simmons 2003; Brennan et al. 2007; Adler 2009; Hotzy and Arnqvist 2009;
Brennan et al. 2010).
Male genital morphology has been implicated in sexual conflict over copulation duration
in garter snakes (King et al. 2009). King et al. (2009) reported that male plains garter
snakes (Thamnophis radix) have large, bilobed hemipenes and prolonged copulation
122
durations ( ̅ = 98 min) compared with those of common garter snakes ( ̅ = 17 min,
Thamnophis sirtalis), which exhibit simple cylindrical hemipenes. Further, female plains
garter snakes exhibited vigorous body-rolling behavior, which shortened copulation
durations significantly. Body rolls were interpreted as female unreceptive behavior in T.
radix, while female T. s. sirtalis did not roll (King et al. 2009). However, sexual conflict
has also been described as a prominent feature in the mating system of the red-sided
garter snake, Thamnophis sirtalis parietalis, (e.g., Shine et al. 2000b; Shine et al. 2004), a
subspecies of the common garter snake. Rolling behavior is common in this subspecies
(Friesen and Mason pers. Obs.), and females may actively prevent sperm from entering
their oviducts during copulation which forces sperm into the copulatory plug (Friesen et
al. in prep.). Copulation duration in T.s. parietalis is not correlated with the number of
sperm inseminated, but is correlated with the mass of the relatively sperm-free caudal
portion of the copulatory plug (Friesen et al. in prep.). King et al. (2009) addressed the
gross morphology of the hemipenes in their analysis (uni-lobed versus bilobed),
however, they did not address the various calyces and spines that adorn the
intromittent organs of male garter snakes, which display so much variation among
closely related species of colubrids that they were used extensively in snake systematics
before the advent of molecular techniques (Cope 1984; Dowling 1967). Dowling (1967)
argued that hemipenial traits had “no obvious correlation with ecology”. Nevertheless,
the arbitrary evolutionary trajectories generated by sexually antagonistic arms races
could explain the great diversity of male and female genitalia among numerous closely
123
related taxa (Hosken and Stockley 2004; but also see Eberhard 1996; 2004). One way to
test adaptive hypotheses of sexual conflict during mating is to manipulate females and
males such that the putative adaptations for conflict are rendered ineffective. For
example, intromittent organ manipulation has been used to test for sexual conflict over
mating in insects by disentangling male and female contributions to copulation duration
(Takami 2003; Rodríguez et al. 2004; Polak and Rashed 2010).
To determine whether plug deposition and copulation duration is a source of conflict in
snakes, we first attempted to show that females actively limit copulation duration by
manipulating the ability of females to resist plug deposition using a local anesthetic
applied to the cloacal region (e.g., Mendonça and Crews 1990, 2001). We then focused
on the role of the basal spine of the male hemipene as a morphological adaptation that
allows males to control plug deposition. The basal spines of the hemipene are the
largest spines (Figure 5.0) and the first to make contact with the female cloaca during
the initial stages of intromission. Males are much less likely to gain intromission with
females when this spine is ablated (Brennan unpublished data). However, many males
gain intromission even without the basal spine, and preliminary evidence suggested that
copulation duration was reduced with spine ablation. To test the role the basal spine
had in plug deposition and to disentangle male and female roles in the control
copulation, we used local anesthetics to limit female control during mating.
124
Figure 5.0: Photograph of the partially everted right hemipene of T.s.parietalis. The basal spine
is indicated by the arrow (scale: each hash mark = 1 mm).
125
5.2
Methods
Female only manipulations and mating trials Spring 2011:
Twenty four seasonal virgin females were collected one day before mating trials (20,
May 2011) from the Inwood quarry den-study site in the Interlake region of Manitoba,
Canada (50° 31.58’N 97°29.71’W), and taken to the Chatfield Research Station 16 km
away. Females were housed in seminatural nylon arenas and were provided water ad
libitum until mating trials. The next day these females were returned to Inwood quarry
for assisted mating trials. Three considerations informed our choice of assisted matings
over natural matings in this study: 1) A preliminary study of basal spine function
demonstrated a significant role for the spine in initiating matings. Over the course of the
trials females would be better able to reject spine ablated males versus controls if they
were allowed to mate naturally. 2) If the level of female control over the initiation of
mating differed between treatments in a way that influenced male-female interactions
during copulation this difference could mask the effect of spine ablation during
copulation. 3) Conflict over copulation and plug deposition, if it occurs at all, would be
most clearly manifest during involuntary matings.
Females randomly assigned to the local anesthesia treatment (N = 12) received two 30µl
bilateral injections of 0.5% Marcaine® (Bupivacaine HCl) directly lateral to the cloaca
between the first and second dorsal scale rows approximately 30 min prior to mating
trials. Control females (N = 12) were treated in exactly the same way as the anesthetized
females except they received injections of saline in place of Marcaine®. All females
126
were marked for identification using a 0.25cm x 0.25cm piece of 3M, Nexcare adhesive
tape fixed to their head with a letter written in permanent marker (Sharpie™).
Immediately after the injections, the females were placed in small arenas (45 cm dia. x
75 cm tall). Thirty minutes was allowed for the anesthetic to take effect, then females
were placed in a natural aggregation of males within the epicenter of the den. Each
female was gently held 1 cm caudal of the cloaca while males courted her. Assisted
matings were conducted as a courting male aligned with the female. We used a blunt
probe to lift the ventral scale that covers the opening to her cloaca. Typically the male
would evert one of his two hemipenes into the female’s cloaca after only a few seconds.
When mating occurred, a stopwatch was started immediately. After 1 min of copulation,
the pair was gently moved to a small circular arena (45 cm dia. x 75 cm tall) where they
were constantly observed until copulation terminated and the duration was recorded (±
10s). The copulatory plug was collected using a blunt probe (Friesen et al. submitted)
and the plug mass was measured (± 0.01g) using a Mettler BB 2400 digital scale.
Male and female manipulations 2012
Hemipene basal spine ablation 2012
Vigorously active courting males (N = 70) were collected one day before the spine
ablation procedure (6, May 2012) from the study site and taken to the Chatfield
Research Station 16 km away. Males were housed in seminatural nylon arenas and were
provided water ad libitum until hemipene spine ablation. Each male of 35 size-matched
pairs was randomly assigned to receive either a basal spine ablation treatment or
127
control treatment. Both treatments consisted of lightly anesthetizing males with 0.0015
µl of 0.5% methohexital sodium/g body mass administered subcutaneously at the
juncture between the dorsal and ventral scales at a distance of approximately 4 cm from
the head of the snake (Preston et al. 2010). Ten minutes after anesthesia was
administered, each male received two 10µl bilateral subcutaneous injections of the local
anesthetic 2% Lidocaine HCl. The injection sites were both between the first and second
dorsal scale rows and four subcaudal scales caudal to the cloaca. Five to ten minutes
after Lidocaine injections, both hemipenes were everted manually by slipping a 2 cm
long piece of 0.9cm diameter latex tubing over the tail and sliding it cranially, just caudal
of the cloaca and, finally, squeezing the tube cranially until the basal spines of both
hemipenes were exposed. Once the basal spines were exposed, a clamp was lightly
secured on the tubing posterior to the cloaca to prevent the hemipenes from inverting
back into their sheaths. The largest basal-most spine was removed from both the left
and right hemipenes of the spine ablated males (N = 35) by clipping it close to its base
with sterilized corneoscleral scissors under a dissecting scope. Control males (N = 35)
were treated in the same way except that the basal spine was touched with the scissors
but not clipped. The hemipenes of all males were swabbed first with reptile ringer’s
solution and then 70% ethyl alcohol before removing the clamp, which allowed the
hemipenes to invert naturally. The snakes were then placed in an aquarium with a
heating pad (30°C) and monitored every 10 minutes for the first hour after the
procedure until righting reflex returned (Preston et al. 2010). Males of this species
128
regain their drive to court very soon even after serious surgeries and show no deficit in
this ability (e.g., Nelson et al. 1987). All of the procedures were completed in one day (7,
May 2012) and all animals survived and engaged in courtship within 24 hours after
recovery.
Female collection and treatments 2012
For three days prior to spine ablation procedures, newly emerged female garter snakes
were collected (N = 100) from the same study site and taken to the Chatfield Research
Station. Females were housed in seminatural nylon arenas and were provided water ad
libitum until mating trials. All females were seasonal virgins as they were collected
immediately upon emergence from winter hibernation.
Mating trials and female treatments 2012
Mating trials were conducted over the course of two days (8-9, May 2012) and were
first seasonal-matings for the all males. The right and left oviductal sphincters of each
female were checked to determine whether they were open or closed using a 20ga.
intubation needle affixed to a 1 ml syringe prior to administering the treatments. Those
females assigned to the local anesthesia treatment received two 30µl bilateral injections
of 0.5% Marcaine® (Bupivacaine HCl) directly lateral to the cloaca between the first and
second dorsal scale rows approximately 30 min prior to mating trials. Control females
were treated in exactly the same was and the anesthetized females except they
received injections of saline in place of Marcaine®. Immediately after injections the
females were placed in small arenas (45 cm dia. x 75 cm tall). Thirty minutes was
129
allowed for the anesthetic to take effect, and then females were placed into larger
1mx1mx1m arenas with males for the mating trails.
Mating trials consisted of randomly assigning four size-matched females to either the
treatment (Marcaine® injected) or control (saline injected) group, such that all four
combinations of male and female treatments had size matched females within them.
Each group of males was randomly assigned (coin flip) to one of two 1mx1mx1m
seminatural arenas; males were allowed to court the females. Thus, each arena
contained 35 males from one of the two male treatment groups; this sex ratio is
common in and around the dens (Joy and Crews 1985; Shine et al. 2006). In this way we
could assure a relatively balanced design in that we could allow matings from each
pairing to occur in roughly the same temporal pattern over the course of the mating
trails. Each female was assisted to mate with one of the males courting her (e.g., Friesen
et al. in prep.), (N = 62) over only two days.
Assisted matings were conducted as those described above, except the matings were
conducted within 1mx1mx1m nylons arenas. The experimenter conducting the assisted
matings was blind to both male and female treatments and moved from one arena to
the other between each mating to ensure both male treatment groups were in lock-step
throughout the trials. Any signs of female resistance and unreceptive behavior (darting
away from courting males and/or body rolls, e.g., King et al. 2009) during courtship and
during the initial stage of copulation were noted. When mating occurred, a stopwatch
130
was started immediately. After 1 min the pair was gently moved to a small circular arena
(45 cm dia. x 75 cm tall) where they were under constant observation. In this way
copulation duration was recorded (± 10s). Each time a male mated, an actively courting
size-matched male from the nearby den replaced the experimental male to maintain the
same density and operational sex ratio within the mating arenas. The replacement male
was marked with a permanent ink marker (Sharpie™, 2 cm on the dorsal stripe) and was
allowed to court but not mate with the experimental females.
Copulatory plug and oviductal sperm collection
Less than 30s after copulation terminated, each female was inspected for a copulatory
plug. Each plug was removed by gently running a blunt probe around the plug to
separate it from the walls of vaginal pouch. Once removed, the plug was placed in a
sealed pre-massed (± 0.001g) 1.5 ml microcentrifuge tube and then placed in a cooler
with ice packs until processing. The oviducts were checked again and the cloaca and
open oviducts were lavaged with 100-500 µl of modified Ham’s F-10 medium and 10
µg/ml of the antibiotic Gentamicin Sulfate (Cat # 99175, Irvine Scientific; e.g., Mattson
et al. 2007) using a 20ga. intubation needle affixed to a 1ml syringe. The fluid from the
vaginal/oviductal-wash would contain any sperm not embedded within the plug and
was subsequently added to a 1.5ml tube.
Plug mass and sperm counts
At the end of each day’s mating trials, each plug mass was determined (± 0.001g) using a
A&D Fx-300i scale, and then placed in 1.5 ml microcentrifuge tubes containing 1ml of
131
modified Ham’s F-10 medium and 10 µg/ml of the antibiotic Gentamicin Sulfate. The
tubes were placed in a refrigerator for two days and were gently agitated three times
daily to aid the liberation of sperm embedded within the plug. The dissolution of the
plug was evidenced by a dense “cloud” of sperm above a much reduced plug. The
approximate sperm concentration of the samples (1µL) was estimated with a
microscope and a counting chamber, and then diluted to facilitate accurate sperm
counts (~ 100 sperm within the counting grid). After dilution, photos of triplicate
samples spread over a Petroff-Hausser sperm cell counter (cat. # 3900, Hausser
Scientific) were taken with an Olympus DP-5 digital camera mounted on an Olympus
CX31 phase contrast compound microscope using the 4x objective. The images were
captured using Cell-Sense software from Olympus and the sperm counts were
conducted using the same software in the lab at Oregon State University. Sperm counts
from the vaginal/oviductal washes were conducted in the same manner.
5.3
Results
Across all treatments excluding seven matings in which plugs were not deposited (2012,
N = 61), plug mass increases with male size (snout to vent length, Msvl; Adj. R2 = 0.109,
F1, 54 =7.629, P = 0.008) Figure 5.1a. Copulation duration was not affected by male size
(Msvl; Adj. R2 = 0.000, F1,60 = 0.325, P = 0.571) Figure 5.1b, or female size (Fsvl; Adj. R2 =
0.000, F1,60 = 0.0105, P = 0.919). Plug mass increased with copulation duration (Adj. R2 =
0.317, F1, 60 = 28.785, P < 0.001) Figure 5.2a; this result is robust to the removal of very
short copulations (≤ 120s) that did not produce plugs, but explained much less of the
132
variation than when the data were included (Adj. R2 = 0.081, P = 0.020). Backward stepwise model selection found a significant effect of a male x female interaction on plug
mass (MxF interaction; Adj. R2 = 0.147, F1, 54 = 10.277, P = 0.002), however, this effect is
confounded by a weak tendency for larger males to mate with larger females (Adj. R 2 =
0.034, F1, 61 = 3.171, P 0.08) and female size alone does not explain plug mass (Fsvl; Adj.
R2 = 0.0152, F1,54 = 1.832, P = 0.182) Figure 5.2b.
133
134
Treatment effects
The data collected from 2011 were analyzed separately from 2012. During 2011,
copulation duration was significantly longer in the Marcaine® treated females than
controls (t22 = 2.374, P = 0.027) Figure 5.3a, but plug mass was not significantly different
(t22 = 0.898, P = 0.379) Figure 5.3b.
In 2012, mean plug mass was not significantly different between anesthetized and
control treated females (t60 = -0.922, P = 0.360). We removed a single extreme
copulation duration that was greater than 75 minutes long from the analysis (hence 59
degrees of freedom rather than 60). Copulation duration was significantly more variable
in spine-ablated males than spine-intact males (Levene’s test; F1, 59 = 5.670, P = 0.042),
but the median time in copula was not different (Mann-Whitney U = 406.000, P =
0.406). Plug mass was also significantly more variable in spine-ablated males than
control spine-intact males (Levene’s test; F 1, 60 = 11.538, P = 0.002) Figure 5.4, but
median plug mass was not different (Mann-Whitney U Statistic= 452.000, P = 0.698).
There was a significant interaction between male size (Msvl) and male x female
treatment (ANOVA F4,54 = 4.056, P = 0.006) Table 5.1 and Figure 5.5. Anesthetized
females tended to mate for longer than controls, but this did not reach significance (t59
= -1.686, P = 0.097; expected given 2011 results, one-tailed P = 0.049). However, if the
effects male treatment and the male x female treatment interaction are included in the
model female treatment has a significant effect on copulation duration Table 5.2 and
Figure 5.6.
135
136
137
138
Oviductal sperm
Most sperm were found in the plugs: median number of sperm in the plug was 3.7x10 7
(SEM ± 5.1x106) and in the oviducts/vaginal pouch 1.0x104 (SEM ± 4.6x105); (Wilcoxon
signed rank test W = 1540, P < 0.001) Figure 5.7a. The condition of the oviductal
sphincter (open versus closed) had a significant effect on the number of sperm collected
from the vaginal pouch and oviducts (Mann-Whitney U = 165, P < 0.001), Figure 5.7b.
Akaike’s AIC model selection was used to determine whether there were significant
treatment effects that explained oviductal sperm counts including female x male
treatment interaction, and oviductal condition (ANOVA F5, 61 = 3.555, P = 0.007) Table
5.3 and Figure 5.8.
139
140
5.4
Discussion
We demonstrate that copulation in red-sided garter snakes is a complex interaction
between males and females. Longer copulation durations result in larger plugs;
however, larger males seem to deposit larger plugs without a concomitant increase in
copulation duration. One explanation for these results is that larger males ejaculate
renal sexual segment RSS fluids at a higher rate than smaller males; these are the fluids
that make up the plug matrix, but do not contain sperm. An alternative explanation is
that females are less resistant to plug deposition when they mate with larger males. In a
previous experiment, we demonstrated that larger males do not inseminate more
sperm than smaller males, but clearly the current experiment demonstrates that they
produce larger plugs without spending more time than smaller males. Plug deposition
likely depends on many vagaries of which we are unaware.
Male size remained a significant predictor of plug mass across all treatments, and
treatment effects on plug mass were unmasked only after accounting for male size.
Overall, spine-ablated males demonstrated less control over plug deposition than spineintact males evidenced by the fact that they showed considerably more variation in plug
mass and copulation duration than spine-intact controls. Most of this variation was
exhibited when males of either treatment mated with anesthetized females. Perhaps
paradoxically, spine-intact males actually deposited smaller plugs when they mated with
anesthetized females compared with spine-ablated males. It is possible that when
Marcaine®-treated females mate, they have no control over the size of their vaginal
141
pouch which is highly muscularized. This is the effect we hoped to achieve with this
treatment. When anesthetized females mated with spine-intact males, these males may
have intromitted their hemipene much deeper into the vaginal pouch thus leaving less
space to fill with plug material. In contrast, spine ablated males were not able to
intromit as deeply and the female’s vaginal pouch was not as constricted due to loss of
muscular control, therefore, these plugs were larger. Spine-ablated males mated with
control females were still unable to penetrate into the vaginal pouch as far as spineintact males, but these females were able to constrict the pouch. One unexamined
assumption implicit in the preceding explanation is that males always benefit from
increased plug mass. As all components of the ejaculate are costly (Dewsbury 1982;
Olsson et al. 1997), there may be an optimal amount of plug material to deposit (maybe
particular to male condition). Further, when females are resistant during copulation
males are unable to disengage. Males may want to prudently adjust their ejaculate
(Wedell et al. 2002), but female resistance during copulation may prevent them from
doing so. The male’s hemipene has many smaller spines, which we did not remove, and
these would have been the sole attachment for the spine-ablated males. It may have
been difficult for the spine-ablated males to disengage these smaller spines. Another
possibility is that intact males interpreted the lack of female resistance as an indicator of
poor female condition. Consequently, intact males may have allocated less of a valuable
resource to a poor quality female that may not give birth or survive.
142
Clearly females exert some control of copulation duration, which does lead to the
deposition of larger plugs as revealed in the analysis of pooled data. In 2011, Marcaine®treated females had significantly longer copulation durations than controls when they
mated with spine-intact males, but plug masses were not different. That year plug mass
was not measured as precisely as in 2012 and we believed that may have accounted for
our inability to detect a difference. In 2012, increased copulation duration was only
exhibited when the females mated with spine-ablated males and the model included
the interaction between male and female treatment. Overall, these results indicate that
females do seem to resist mating and limit copulation duration, but the longer
copulation durations do not account for increases in plug mass across treatments.
Most of the sperm inseminated is contained in the copulatory plug rather than in the
oviducts as previously thought (Devine 1975). An anatomical explanation for our
observation is that there are oviductal sphincters which may close during copulation and
thus prevent sperm from entering the oviducts causing it to swirl into the cranial to
medial portions of the plug matrix. We had hoped to manipulate female control of these
sphincters, but whether they were opened or closed was not affected by local
anesthesia. Regardless, our results indicate that the sphincters prevent sperm from
entering the oviducts during copulation, but even if they are open, contraction of the
vaginal pouch prevents sperm from entering the oviducts. Further, vaginal contraction is
not required if the basal spine has been removed. Thus, there is a three-way interaction
between the three conditions which must be met in order for sperm to enter the
143
oviducts: the sphincters must be open, the vaginal pouch must not contract, and the
male must have an intact basal spine. It is unclear why females would tighten vaginal
pouch during copulation, but perhaps it is to prevent bacteria or other foreign materials
from entering her oviducts. It is also possible that vaginal constriction is a way to limit
insemination by unwanted males when copulation cannot be avoided.
Twenty-five percent of snake species do not have hemipenial spines (Olsson and
Madsen 1998), so these spines are not universally required for copulation. Most snakes
have much longer copulation durations than garter snakes (Olsson and Madsen 1998),
and red- sided garter snakes have the shortest copulations in the genus Thamnophis
(King et al. 2009). Our study suggests the basal spines are important in overcoming
female resistance and control over mating during plug deposition. The spines, then, may
be adaptations for quickly depositing a plug to reduce mating opportunity costs of
prolonged copulations as seen in other snake species (Shine et al. 2000a; Friesen et al. in
prep.). From the female perspective, within the large mating aggregations of red-sided
garter snakes of Manitoba, female precopulatory choice is likely to be limited. Females
may use copulation to assess their copulation partner; part of the resolution of conflict
may be to limit the number of sperm he inseminates and the size of the copulatory plug.
Clearly, it would be easier to reject the male before he gains intromission. The basal
spine is the male’s solution to prevent females from rejecting intromission (Brennan
unpublished data), but the spine also plays a role in insemination and possibly
disengagement of copulation. One further study would be to assess paternity (e.g.,
144
Polak and Rashed 2009) among these same experimental groups in order to assess how
sperm transport, subsequent to mating, is affected by the vaginal pouch and the basal
spine. This work provides a useful inroad into further research of sexual conflict during
copulation in this and many other species.
145
THESIS CONCLUSION: FUTURE DIRECTIONS FOR RESEARCH
Red-sided garter snakes are an amazing model system in which to study mating system
evolution and postcopulatory selection. We have provided a descriptive foundation for many
further investigations.
Retiles in general, and snakes in particular are unique in the ubiquity of stored sperm usage. The
theoretical framework of sperm competition has yet to fully incorporate adaptations for longterm sperm storage; however, with mounting empirical evidence gathered from reptile taxa, the
theoreticians are sure to take note.
Although this body of work does not directly assess whether paternity assigned to stored sperm
originates from autumnal matings or vernal matings from the previous year, we have provided
convincing evidence of the prevalence, and thus, importance of stored sperm. Stored sperm
makes up a considerable proportion of paternity in red-sided garter snakes. Whether this has
been a selective force on sperm viability and on female processes that sustain, and perhaps
select sperm, deserves further attention. We examined only two sperm traits, mobility and
numbers, but the morphology of sperm is important to velocity and viability. The sperm of redsided garter snakes have long mid-pieces which have been correlated with viability in other
species, and this to should be investigated. Intra- and interspecific comparisons of various sperm
traits my identify species and populations in which postcopulatory selection is strong and
provide corroborative evidence of its strength in the populations we are currently investigating.
Paternity analysis of natural second matings will shed illuminate whether females bias paternity
in response to unwanted matings. The natural arenas we designed and may be invaluable to the
collection of what are elusive events in a female garter snake’s life once she leaves the den. I am
146
sure many other experiments can be designed that benefit from a more natural setting for
reproductive behavior.
There are many populations of red-sided garter snakes and subspecies of T. sirtalis across North
America with variable aggregation densities, and probably concomitant differences levels of
postcopulatory selection. Studies of the determinants of fertilization success in these
populations and subspecies would provide conclusive tests of our proposed hypothesis that
mating aggregations densities are at the root of the gametic asymmetry we uncovered. Further,
a survey of copulation duration and its correlation with occurrence of copulatory plugs in snakes
would yield insight into the evolutionary significance of plugs and which functions evolved first.
While we found corroborative evidence for our results in the smaller testes and ejaculate size of
Snake Island males; further investigations of the effect of male size within the Snake Island
population should include sperm counts of ejaculates from matings with Snake Island females. If
male size has no effect in those crosses, and yet the male size advantage is reaffirmed through
paternity analysis, then this is strongly suggestive of female mediated processes in which
precopulatory male size advantages are reinforced through postcopulatory selection via cryptic
female choice.
Sexual conflict over copulation seems manifest in the Inwood population. If there is in fact
conflict over plug deposition as the studies presented here indicate, then females may have
enzymes that aid them in quickly dissolving the plug. Again between-population comparisons of
these putative proteins may reveal the signature of rapid evolution of these proteins as has
been discovered in other reproductive proteins. Female control of sperm transport also may be
a fruitful area to explore with regards to sexual conflict and the genital manipulations we have
147
pioneered in these studies. Further, vasectomies and ureter ligations techniques could be used
to isolation the putative proteins or other substances (e.g., sperm, or prostaglandins?), which
may affect female remating rates.
148
BIBLIOGRAPHY
Aleksiuk M, Gregory PT. 1974. Regulation of seasonal mating behavior in Thamnophis sirtalis
parietalis. Copeia 1974: 681-689.
Althaus S, Jacob A, Graber W, Hofer D, Nentwig W, Kropf C. 2010. A double role of sperm in
scorpions: The mating plug of Euscorpius italicus (Scorpiones: Euscorpiidae) consists of
sperm. Journal of Morphology 271: 383-393.
Andersson M, Simmons LW. 2006. Sexual selection and mate choice. Trends in Ecology &
Evolution 21: 296-302.
Andersson MB. 1994. Sexual Selection. Princeton University Press: Princeton, N. J.
Arnold SJ, Duvall D. 1994. Animal mating systems: A synthesis based on selection theory.
American Naturalist 143: 317-348.
Arnold SJ, Verrell PA, Tilley SG. 1996. The evolution of asymmetry in sexual isolation: a model
and a test case. Evolution: 1024-1033.
Arnqvist G, Edvardsson M, Friberg U, Nilsson T. 2000. Sexual conflict promotes speciation in
insects. Proceedings of the National Academy of Sciences 97: 10460-10464.
Arnqvist G, Nilsson T. 2000. The evolution of polyandry: multiple mating and female fitness in
insects. Animal Behaviour 60: 145-164.
Arnqvist G, Rowe L. 2005. Sexual Conflict. Princeton University Press: Princeton, New Jersey.
Asdell SA. 1946. Patterns of Mammalian Reproduction. Comstock Publishing Co.: Ithaca, N.Y.
Birkhead T. 1998. Sperm competition in birds: mechanisms and function. In: Birkhead T and
Møller AP, eds. Sperm competition and sexual selection. San Diego, CA: Academic Press.
579-622.
Birkhead T, Hosken D, Pitnick S. 2008. Sperm biology: an evolutionary perspective. Academic
Press.
Birkhead T, Martinez J, Burke T, Froman D. 1999. Sperm mobility determines the outcome of
sperm competition in the domestic fowl. Proceedings of the Royal Society of London.
Series B: Biological Sciences 266: 1759-1764.
Birkhead TR, Brillard JP. 2007. Reproductive isolation in birds: postcopulatory prezygotic
barriers. Trends in Ecology & Evolution 22: 266-272.
Birkhead TR, Møller AP. 1998. Sperm competition and sexual selection. Academic Press: San
Diego.
Birkhead TR, Pizzari T. 2002. Postcopulatory Sexual Selection. Nature Reviews Genetics 3: 262273.
Bishop JE. 1959. A histological and histochemical study of the kidney tubule of the common
garter snake, Thamnophis sirtalis, with special reference to the sexual segment in the
male. Journal of Morphology 104: 307-357.
Bissoondath C, Wiklund C. 1996. Effect of male mating history and body size on ejaculate size
and quality in two polyandrous butterflies, Pieris napi and Pieris rapae (Lepidoptera:
Pieridae). Functional Ecology 10: 457-464.
Blanchard FC. 1943. A test of fecundity of the garter snake Thamnophis sirtalis sirtalis (Linnaeus)
in the year following the year of insemination. Michigan Academy of Sciences, Arts and
Letters 28: 313–316.
Blanchard FN, Blanchard FC. 1941. The inheritance of melanism in the garter snake Thamnophis
sirtalis sirtalis (Linnaeus), and some evidence of effective autumn mating. . Michigan
Academy of Sciences, Arts and Letters 26: 177-193.
149
Blanchard FN, Finster EB. 1933. A method of marking living snakes for future recognition, with a
discussion of some problems and results. Ecology 14: 334-347.
Blandau RJ. 1945. On the factors involved in sperm transport through the cervix uteri of the
albino rat. American Journal of Anatomy 77: 253-272.
Bleu J, Bessa-Gomes C, Laloi D. 2012. Evolution of female choosiness and mating frequency:
effects of mating cost, density and sex ratio. Animal Behaviour 83: 131-136.
Booth W, Smith CF, Eskridge PH, Hoss SK, Mendelson JR, Schuett GW. 2012. Facultative
parthenogenesis discovered in wild vertebrates. Biology Letters.
Boschetto C, Gasparini C, Pilastro A. 2011. Sperm number and velocity affect sperm
competition success in the guppy (Poecilia reticulata). Behavioral Ecology and
Sociobiology 65: 813-821.
Brennan PL, Prum RO, McCracken KG, Sorenson MD, Wilson RE, Birkhead TR. 2007.
Coevolution of male and female genital morphology in waterfowl. PLoS ONE 2: e418.
Brennan PLR, Clark CJ, Prum RO. 2010. Explosive eversion and functional morphology of the
duck penis supports sexual conflict in waterfowl genitalia. Proceedings of the Royal
Society B: Biological Sciences 277: 1309-1314.
Burtner H, Floyd A, Longley J. 1965. Histochemistry of the “sexual segment” granules of the
male rattlesnake kidney. Journal of Morphology 116: 189-195.
Casselman SJ, Schulte-Hostedde AI, Montgomerie R. 2006. Sperm quality influences male
fertilization success in walleye (Sander vitreus). Canadian Journal of Fisheries and
Aquatic Sciences 63: 2119-2125.
Clutton-Brock T, Parker G. 1995. Sexual coercion in animal societies. Animal Behaviour.
Collins A, Williams V, Evans J. 2004. Sperm storage and antioxidative enzyme expression in the
honey bee, Apis mellifera. Insect molecular biology 13: 141-146.
Coyne J, Orr H. 2004. Speciation. Sinauer Sunderland, MA.
Crews D, Camazine B, Diamond M, Mason R, Tokarz RR, Garstka WR. 1984. Hormonal
independence of courtship behavior in the male garter snake. Hormones and Behavior
18: 29-41.
Darwin C. I87I. The Descent of Man and Selection in Relation to Sex. D. Appleton and Company:
New York, NY.
Davies NB. 1983. Polyandry, cloaca-pecking and sperm competition in dunnocks. Nature 302:
334-336.
Devine MC. 1975. Copulatory plugs in snakes: Enforced chastity. Science 187: 844-845.
Devine MC. 1977. Copulatory plugs, restricted mating opportunities and reproductive
competition among male garter snakes Nature. 345-346.
Devine MC. 1984. Potential for sperm competition in reptiles: behavioral and physiological
consequences. In: Smith RL, ed. Sperm Competition and the Evolution of Animal Mating
Systems. 509–521.
Dewsbury DA. 1982. Ejaculate Cost and Male Choice. American Naturalist 119: 601-610.
Dewsbury DA. 1988. A test of the role of copulatory plugs in sperm competition in deer mice
(Peromyscus maniculatus). Journal of Mammalogy 69: 854-857.
DiLeo MF, Row JR, Lougheed SC. 2010. Discordant patterns of population structure for two co‐
distributed snake species across a fragmented Ontario landscape. Diversity and
Distributions 16: 571-581.
150
Dixon SM, Coyne JA, Noor MAF. 2003. The evolution of conspecific sperm precedence in
Drosophila. Molecular Ecology 12: 1179-1184.
Dowling HG. 1967. Hemipenes and other characters in colubrid classification. Herpetologica 23:
138-142.
Dziminski MA, Roberts JD, Beveridge M, Simmons LW. 2009. Sperm competitiveness in frogs:
slow and steady wins the race. Proceedings of the Royal Society B: Biological Sciences
276: 3955-3961.
Dziuk PJ. 1996. Factors that influence the proportion of offspring sired by a male following
heterospermic insemination. Animal Reproduction Science 43: 65-88.
Eady PE. 2006. Postcopulatory, prezygotic reproductive isolation. Journal of Zoology 253: 47-52.
Eberhard W. 2004. Rapid divergent evolution of sexual morphology: Comparative tests of
antagonistic coevolution and traditional female choice. Evolution 58: 1947-1970.
Eberhard WG. 1996. Female Control: Sexual Selection by Cryptic Female Choice. Princeton
University Press: Princeton, New Jersey.
Elgee K, Evans J, Ramnarine I, Rush S, Pitcher T. 2010. Geographic variation in sperm traits
reflects predation risk and natural rates of multiple paternity in the guppy. Journal of
Evolutionary Biology 23: 1331-1338.
Emlen ST, Oring LW. 1977. Ecology, sexual selection, and the evolution of mating systems.
Science 197: 215-223.
Evans JP, Magurran AE. 2001. Patterns of sperm precedence and predictors of paternity in the
Trinidadian guppy. Proceedings of the Royal Society of London. Series B: Biological
Sciences 268: 719-724.
Fahrig BM, Mitchell MA, Eilts BE, Paccamonti DL. 2007. Characterization and cooled storage of
semen from corn snakes (Elaphe guttata). Journal of Zoo and Wildlife Medicine 38: 7-12.
Fitch HS. 1965. An ecological study of the garter snake, Thamnophis sirtalis. University of Kansas
Publications of the Museum of Natural History 15: 493-564.
Fox CW, Rauter CM. 2003. Bet-hedging and the evolution of multiple mating. Evolutionary
Ecology Research 5: 273-286.
Fox H. 1977. The urogenital system of reptiles. Biology of the Reptilia 6: 1-157.
Fox W. 1956. Seminal receptacles of snakes. The Anatomical record 124: 519-539.
Froman D, Feltmann A. 2000. Sperm mobility: Phenotype in roosters (Gallus domesticus)
determined by concentration of motile sperm and straight line velocity. Biology of
Reproduction 62: 303-309.
Froman DP, Feltmann AJ, Rhoads ML, Kirby JD. 1999. Sperm mobility: A primary determinant of
fertility in the domestic fowl (Gallus domesticus). Biology of Reproduction 61: 400.
Froman DP, McLean DJ. 1996. Objective Measurement of Sperm Motility Based Upon Sperm
Penetration of Accudenz®. Poultry science 75: 776-784.
Gabor CR, Halliday TR. 1997. Sequential mate choice by multiply mating smooth newts: females
become more choosy. Behavioral ecology : official journal of the International Society
for Behavioral Ecology. 8: 162.
Gage MJG, Morrow EH. 2003. Experimental Evidence for the Evolution of Numerous, Tiny
Sperm via Sperm Competition. Current Biology 13: 754-757.
151
Garner TWJ, Gregory PT, McCracken GF, Burghardt GM, Koop BF, McLain SE, Nelson RJ. 2002.
Geographic variation of multiple paternity in the common garter snake (Thamnophis
sirtalis). Copeia 2002: 15-23.
Garner TWJ, Larsen KW. 2005. Multiple paternity in the western terrestrial garter snake,
Thamnophis elegans. Canadian Journal of Zoology 83: 656-663.
Garner TWJ, Pearman PB, Gregory PT, Tomio G, Wischniowski SG, Hosken DJ. 2004.
Microsatellite markers developed from Thamnophis elegans and Thamnophis sirtalis
and their utility in three species of garter snakes. Molecular Ecology Notes 4: 369-371.
Garstka WR, Camazine B, Crews D. 1982. Interactions of behavior and physiology during the
annual reproductive cycle of the red-sided garter snake (Thamnophis sirtalis parietalis).
Herpetologica: 104-123.
Gavrilets S. 2000. Rapid evolution of reproductive barriers driven by sexual conflict. Nature 403:
886-889.
Gavrilets S, Hayashi T. 2005. Speciation and sexual conflict. Evolutionary Ecology 19: 167-198.
Geyer LB, Palumbi SR. 2005. Conspecific sperm precedence in two species of tropical sea
urchins. Evolution 59: 97-105.
Gibson AR, Falls JB. 1975. Evidence for multiple insemination in the common garter snake,
Thamnophis sirtalis. Canadian Journal of Zoology 53: 1362-1368.
Gould SJ, Lewontin RC. 1979. The spandrels of San Marco and the Panglossian paradigm: a
critique of the adaptationist programme. Proceedings of the Royal Society of London.
Series B. Biological Sciences 205: 581-598.
Gowaty P. 1997. Sexual dialectics, sexual selection, and variation in reproductive behavior
Feminism and evolutionary biology: Boundaries, intersections and frontiers. 351–384.
Gregory P. 2006. Influence of income and capital on reproduction in a viviparous snake: direct
and indirect effects. Journal of Zoology 270: 414-419.
Gregory PT. 1974. Patterns of spring emergence of the red-sided garter snake (Thamnophis
sirtalis parietalis) in the Interlake region of Manitoba. Canadian Journal of Zoology 52:
1063–1069.
Gregory PT, Stewart KW. 1975. Long-distance dispersal and feeding strategy of the red-sided
garter snake (Thamnophis sirtalis parietalis) in the Interlake of Manitoba Can. J. Zool.
238-245.
Halliday T, Arnold SJ. 1987. Multiple mating by females: A perspective from quantitative
genetics. Animal Behaviour 35: 939-941.
Happ GM. 1969. Multiple sex pheromones of the mealworm beetle, Tenebrio molitor Nature
222: 180-181.
Hawley AWL, Aleksiuk M. 1975. Thermal regulation of spring mating behavior in the red-sided
garter snake (Thamnophis sirtalis parietalis). Canadian Journal of Zoology 53: 768-776.
Head ML, Brooks R. 2006. Sexual coercion and the opportunity for sexual selection in guppies.
Animal Behaviour 71: 515-522.
Hoffman LH, Wimsatt WA. 1972. Histochemical and electron microscopic observations on the
sperm receptacles in the garter snake oviduct. American Journal of Anatomy 134: 71-95.
Höggren M, Tegelström H. 2002. Genetic evidence for firstmale mating advantage in the adder
(Vipera berus). In: Schuett G, Höggren M, Douglas M and Greene H, eds. Biology of the
Vipers. Utah: Eagle Mountain. 235-242.
152
Hosken D, Blanckenhorn W, Garner T. 2002. Heteropopulation males have a fertilization
advantage during sperm competition in the yellow dung fly (Scathophaga stercoraria).
Proceedings of the Royal Society of London. Series B: Biological Sciences 269: 1701-1707.
Hosken D, Garner T, Tregenza T, Wedell N, Ward P. 2003. Superior sperm competitors sire
higher–quality young. Proceedings of the Royal Society of London. Series B: Biological
Sciences 270: 1933-1938.
Hosken D, Ward P. 2008. Experimental evidence for testis size evolution via sperm competition.
Ecology Letters 4: 10-13.
Hosken DJ, Stockley P. 2003. Benefits of polyandry: A life history perspective. Evolutionay
Biology 33: 173-194.
Hosken DJ, Stockley P. 2004. Sexual selection and genital evolution. Trends in Ecology &
Evolution 19: 87-93.
Hotzy C, Arnqvist G. 2009. Sperm competition favors harmful males in seed beetles. Current
Biology 19: 404-407.
House CM, Simmons LW. 2003. Genital morphology and fertilization success in the dung beetle
Onthophagus taurus: An example of sexually selected male genitalia. Proceedings of the
Royal Society of London Series B-Biological Sciences 270: 447-455.
Howard DJ. 1999. Conspecific sperm and pollen precedence and speciation. Annual Review of
Ecology and Systematics 30: 109-132.
Jennions MD, Petrie M. 2000. Why do females mate multiply? A review of the genetic benefits.
Biological Reviews 75: 21-64.
Jia Z, Duan E, Jiang Z, Wang Z. 2002. Copulatory plugs in masked palm civets: Prevention of
semen leakage, sperm storage, or chastity enhancement? Journal of Mammalogy 83:
1035-1038.
Johnson L, Welch G. 1999. Sex preselection: high-speed flow cytometric sorting of X and Y
sperm for maximum efficiency. Theriogenology 52: 1323-1341.
Jones AG, Arguello JR, Arnold SJ. 2004. Molecular parentage analysis in experimental newt
populations: the response of mating system measures to variation in the operational sex
ratio. The American Naturalist 164: 444-456.
Joy JE, Crews D. 1985. Social dynamics of group courtship behavior in male red-sided garter
snakes (Thamnophis sirtalis parietalis). Journal of Comparative Psychology 99: 145-149.
Kalinowski S, Taper M, Marshall T. 2007. Revising how the computer program cervus
accommodates genotyping error increases success in paternity assignment. Molecular
Ecology 16: 1099-1106.
Kelly CD, Jennions MD. 2011. Sexual selection and sperm quantity: meta‐analyses of strategic
ejaculation. Biological Reviews 86: 863-884.
King RB, Jadin RC, Grue M, Walley HD. 2009. Behavioural correlates with hemipenis
morphology in New World natricine snakes. Biological Journal of the Linnean Society 98:
110-120.
King RB, Milstead WB, Gibbs HL, Prosser MR, Burghardt GM, McCracken GF. 2001. Application
of microsatellite DNA markers to discriminate between maternal and genetic effects on
scalation and behavior in multiply-sired garter snake litters. Canadian Journal of Zoology
79: 121-128.
Kitchell JF. 1969. Thermophilic and thermophobic responses of snakes in a thermal gradient.
Copeia 1969: 189-191.
153
Klug H, Heuschele J, Jennions M, Kokko H. 2010. The mismeasurement of sexual selection.
Journal of Evolutionary Biology 23: 447-462.
Kokko H, Mappes J. 2005. Sexual selection when fertilization is not guaranteed. Evolution 59:
1876-1885.
Koprowski JL. 1992. Removal of copulatory plugs by female tree squirrels. Journal of
Mammalogy 73: 572-576.
Krohmer RW. 2004. Variation in seasonal ultrastructure of sexual granules in the renal sexual
segment of the northern water snake, Nerodia sipedon sipedon. Journal of Morphology
261: 70-80.
Krohmer RW, Grassman M, Crews D. 1987. Annual reproductive cycle in the male red-sided
garter snake, Thamnophis sirtalis parietalis: field and laboratory studies. Gen Comp
Endocrinol 68: 64-75.
Lande R. 1981. Models of Speciation by Sexual Selection on Polygenic Traits. Proceedings of the
National Academy of Sciences 78: 3721-3725.
Larsen KW, Gregory PT, Antoniak R. 1993. Reproductive ecology of the common garter snake,
Thamnophis sirtalis, at the northern Limit of its range. American Midland Naturalist 129:
336-345.
Larson E, Hume G, ANDRéS J, Harrison R. 2012. Post‐mating prezygotic barriers to gene
exchange between hybridizing field crickets. Journal of Evolutionary Biology.
LeMaster MP, Mason RT. 2002. Variation in a female sexual attractiveness pheromone controls
male mate choice in garter snakes. J Chem Ecol 28: 1269-1285.
Lemaster MP, Mason RT. 2003. Pheromonally mediated sexual isolation among denning
populations of red-sided garter snakes, Thamnophis sirtalis parietalis. J Chem Ecol 29:
1027-1043.
Leuckart R. 1847. Zur Morphologie und Anatomie der Geschlechtsorgane. Vandenhoeck und
Ruprecht: Göttingen.
Long JA, Evans HML. 1922. The oestrous cycle in the rat and its associated phenomena.
University of California press: Berkeley, CA.
Lorch P. 2002. Understanding reversals in the relative strength of sexual selection on males and
females: A role for sperm competition? The American Naturalist 159: 645-657.
Lüpold S, Manier MK, Ala-Honkola O, Belote JM, Pitnick S. 2011. Male Drosophila melanogaster
adjust ejaculate size based on female mating status, fecundity, and age. Behavioral
Ecology 22: 184-191.
Lutterschmidt DI, LeMaster MP, Mason RT. 2006. Minimal overwintering temperatures of redsided garter snakes (Thamnophis sirtalis parietalis): a possible cue for emergence?
Canadian Journal of Zoology 84: 771-777.
Maan ME, Seehausen O. 2011. Ecology, sexual selection and speciation. Ecology Letters 14: 591602.
Madsen T, Shine R, Loman J, HÃ¥kansson T. 1992. Why do female adders copulate so
frequently? Nature. 440-441.
Marshall J, Arnold M, Howard D. 2002. Reinforcement: the road not taken. Trends in Ecology &
Evolution 17: 558-563.
Martan J, Shepherd BA. 1976. The role of the copulatory plug in reproduction of the guinea pig.
Journal of Experimental Zoology 196: 79-83.
154
Martín‐Coello J, Benavent‐Corai J, Roldan ERS, Gomendio M. 2009. Sperm competition
promotes asymmetries in reproductive barriers between closely related species.
Evolution 63: 613-623.
Martin NH, Willis JH. 2007. Ecological divergence associated with mating system causes nearly
complete reproductive isolation between sympatric Mimulus species. Evolution 61: 6882.
Martin OY, Hosken DJ. 2003. The evolution of reproductive isolation through sexual conflict.
Nature 423: 979-982.
Mason R, MacMillan S, Whittier J, Krohmer R, Koonz W. 1991. Thamnophis sirtalis parietalis
(red-sided garter snake) Population morph variation. Herpetological Review 22: 61.
Mason RT. 1993. Chemical ecology of the red-sided garter snake, Thamnophis sirtalis parietalis.
Brain Behavior and Evolution 41: 261-261.
Mason RT, Crews D. 1985. Female mimicry in garter snakes. Nature 316: 59-60.
Matson PL, Myssonski K, Yovich S, Morrison L, Irving J, Bakos HW, Australia W, IVF S, Wales
NS. 2010. The density of human semen and the validation of weight as an indicator of
volume: A multicentre study. Reproductive Biology 10: 141-153.
Mattson KJ, Vries AD, McGuire SM, Krebs J, Louis EE, Loskutoff NM. 2007. Successful artificial
insemination in the corn snake, Elaphe gutatta, using fresh and cooled semen. Zoo
Biology 26: 363-369.
McCracken GF, Burghardt GM, Houts SE. 1999. Microsatellite markers and multiple paternity in
the garter snake Thamnophis sirtalis. Molecular Ecology 8: 1475-1479.
Mendonça M, Crews D. 2001. Control of attractivity and receptivity in female red-sided garter
snakes. Hormones and Behavior 40: 43-50.
Mendonça MT, Crews D. 1990. Mating-induced ovarian recrudescence in the red-sided garter
snake. Journal of Comparative Physiology A: Sensory, Neural, and Behavioral Physiology
166: 629-632.
Michener G. 1984. Age, sex, and species differences in the annual cycles of ground-dwelling
sciurids: implications for sociality. In: Murie J and Michener G, eds. The Biology of the
Ground-Dwelling Squirrels: Annual Cycles, Behavioral Ecology, and Sociality. Lincoln,
Nebraska: University of Nebraska Press. 81-107.
Miller GT, Pitnick S. 2002. Sperm-female coevolution in Drosophila. Science 298: 1230-1233.
Mooi RD, Wiens JP, Casper GS. 2011. Extreme color variation within populations of the common
gartersnake, Thamnophis sirtalis, in Central North America, with implications for
subspecies status. Copeia 2011: 187-200.
Moreira PL, Birkhead TR. 2003. Copulatory plugs in the Iberian rock lizard do not prevent
insemination by rival males. Functional Ecology 17: 796-802.
Moreira PL, Nunes VL, Martín J, Paulo OS. 2007. Copulatory plugs do not assure high first male
fertilisation success: Sperm displacement in a lizard. Behavioral Ecology and
Sociobiology 62: 281-288.
Nelson RJ, Mason RT, Krohmer RW, Crews D. 1987. Pinealectomy blocks vernal courtship
behavior in red-sided garter snakes. Physiology & Behavior 39: 231-233.
Nilson G, Andrén C. 1982. Function of renal sex secretion and male hierarchy in the adder,
Vipera berus, during reproduction. Hormones and Behavior 16: 404-413.
O’Donnell RP, Shine R, Mason RT. 2004. Seasonal anorexia in the male red-sided garter snake,
Thamnophis sirtalis parietalis. Behavioral Ecology and Sociobiology 56: 413-419.
155
Oku K, Kitsunezuka K. 2011. Effects of male mating interval on spermatophore formation,
transfer, and subsequent female receptivity and fecundity in the sorghum plant bug,
Stenotus rubrovittatus. Entomologia Experimentalis Et Applicata 140: 134-138.
Olsson M. 2001. 'Voyeurism'prolongs copulation in the dragon lizard Ctenophorus fordi.
Behavioral Ecology and Sociobiology 50: 378-381.
Olsson M, Gullberg A, Tegelstrӧm H. 1994. Sperm competition in the sand lizard, Lacerta agilis.
Animal Behaviour 48: 193-200.
Olsson M, Madsen T. 1998. Sexual selection and sperm competition in reptiles. In: Birkhead TR
and Moller AP, eds. Sperm competition and sexual selection. San Diego, CA: Academic
Press. 503-578.
Olsson M, Madsen T, Shine R. 1997. Is sperm really so cheap? Costs of reproduction in male
adders, Vipera berus. Proceedings of the Royal Society of London. Series B: Biological
Sciences 264: 455-459.
Olsson M, Shine R, Madsen T, Gullberg A, Tegelstroem H. 1996. Sperm selection by females.
Nature 383: 585-585.
Parker G. 1990. Sperm competition games: raffles and roles. Proceedings of the Royal Society of
London. Series B: Biological Sciences 242: 120-126.
Parker G. 1998a. Sperm competition and the evolution of ejaculates: towards a theory base.
Sperm competition and sexual selection 3: 54.
Parker GA. 1970a. The reproductive behavior and the nature of sexual selection in Scatophaga
stercoraria L. (Diptera: Scatophagidae). VII. The origin and evolution of the passive
phase. Evolution 24: 774-788.
Parker GA. 1970b. Sperm competition and its evolutionary consequences in the insects.
Biological Reviews 45: 525-567.
Parker GA. 1979. Sexual selection and sexual conflict. In: Blum MS and Blum NA, eds. Sexual
selection and reproductive competition in insects B2 - Sexual selection and reproductive
competition in insects. New York: Academic Press. 123-163.
Parker GA. 1984. Sperm competition and the evolution of animal mating strategies. In: Smith RL,
ed. Sperm Competition and the Evolution of Animal Mating Systems. New York:
Academic Press. 1-60.
Parker GA. 1998b. Sperm competition and the evolution of ejaculates: towards a theory base.
In: Birkhead TR and Møller AP, eds. Sperm competition and sexual selection. San Deigo
London: Academic Press. 3-49.
Parker GA, Partridge L. 1998. Sexual conflict and speciation. Philosophical Transactions of the
Royal Society B: Biological Sciences 353: 261-274.
Parker GA, Pizzari T. 2010. Sperm competition and ejaculate economics. Biological Reviews 85:
897-934.
Pearse DE, Janzen FJ, Avise JC. 2001. Genetic markers substantiate long-term storage and
utilization of sperm by female painted turtles. Heredity 86: 378-384.
Pilastro A, Simonato M, Bisazza A, Evans JP. 2004. Cryptic female preference for colorful males
in guppies. Evolution International Journal of Organic Evolution 58: 665-669.
Pitcher TE, Neff BD, Rodd FH, Rowe L. 2003. Multiple mating and sequential mate choice in
guppies: females trade up. Proceedings: Biological Sciences 270: 1623-1629.
156
Pitnick S, Brown WD. 2000. Criteria for Demonstrating Female Sperm Choice. Evolution 54:
1052-1056.
Pitnick S, Wolfner MF, Suarez SS. 2009. Ejaculate-female and sperm-female interactions. Sperm
Biology: An Evolutionary Perspective: 247-304.
Pizzari T, Birkhead TR. 2000. Female feral fowl eject sperm of subdominant males. Nature 405:
787-789.
Pizzari T, Worley K, Burke T, Froman DP. 2008. Sperm competition dynamics: Ejaculate
fertilising efficiency changes differentially with time. BMC Evolutionary Biology 8: 332.
Placyk Jr JS, Burghardt GM, Small RL, King RB, Casper GS, Robinson JW. 2007. Post-glacial
recolonization of the Great Lakes region by the common gartersnake (Thamnophis
sirtalis) inferred from mtDNA sequences. Molecular Phylogenetics and Evolution 43:
452-467.
Poiani A. 2006. Complexity of seminal fluid: a review. Behavioral Ecology and Sociobiology 60:
289-310.
Polak M, Rashed A. 2010. Microscale laser surgery reveals adaptive function of male
intromittent genitalia. Proceedings of the Royal Society B: Biological Sciences 277: 13711376.
Preston DL, Mosley CAE, Mason RT. 2010. Sources of variability in recovery time from
Methohexital Sodium anesthesia in snakes. Copeia 2010: 496-501.
Prosser MR, Gibbs HL, Weatherhead PJ. 1999. Microgeographic population genetic structure in
the northern water snake, Nerodia sipedon sipedon detected using microsatellite DNA
loci. Molecular Ecology 8: 329-333.
Ramm SA, Parker GA, Stockley P. 2005. Sperm competition and the evolution of male
reproductive anatomy in rodents. Proceedings of the Royal Society B: Biological Sciences
272: 949-955.
Reinhold K, Kurtz J, Engqvist L. 2002. Cryptic male choice: sperm allocation strategies when
female quality varies. Journal of Evolutionary Biology 15: 201-209.
Ritchie MG. 2007. Sexual selection and speciation. Annu. Rev. Ecol. Evol. Syst. 38: 79-102.
Rivera AC, Andrés J. 2002. Male coercion and convenience polyandry in a calopterygid
damselfly. Journal of Insect Science 2.
Robinson T, Johnson NA, Wade MJ. 1994. Postcopulatory, prezygotic isolation: intraspecific and
interspecific sperm precedence in Tribolium spp., flour beetles. Heredity 73: 155.
Rodríguez V, Windsor D, Eberhard W. 2004. Tortoise beetle genitalia and demonstrations of a
sexually selected advantage for flagellum length in Chelymorpha alternans
(Chrysomelidae, Cassidini, Stolaini). New developments in the biology of Chrysomelidae
2004: 739-748.
Ross P, Crews D. 1978. Stimuli Influencing Mating Behavior in the Garter Snake, Thamnophis
radix. Behavioral Ecology and Sociobiology 4: 133-142.
Ross P, Jr., Crews D. 1977. Influence of the seminal plug on mating behaviour in the garter
snake. Nature 267: 344-345.
Rye LA. 2000. Analysis of areas of intergradation between described subspecies of the common
garter snake, Thamnophis sirtalis in Canada. PhD Thesis The University of Guelph.
Schuett GW, Gillingham JC. 1986. Sperm storage and multiple paternity in the copperhead,
Agkistrodon contortrix. Copeia 1986: 807-811.
157
Schulte-Hostedde A, Montgomerie R. 2006. Intraspecific variation in ejaculate traits of the
northern watersnake (Nerodia sipedon). Journal of Zoology 270: 147-152.
Schwartz JM, McCracken GF, Burghardt GM. 1989. Multiple paternity in wild populations of the
garter snake, Thamnophis sirtalis. Behavioral Ecology and Sociobiology 25: 269-273.
Shine R. 2000. Chastity belts in gartersnakes: The functional significance of mating plugs.
Biological Journal of the Linnean Society 70: 377-390.
Shine R. 2003. Reproductive strategies in snakes. Proceedings of the Royal Society of London.
Series B: Biological Sciences 270: 995-1004.
Shine R, Elphick MJ, Harlow PS, Moore IT, LeMaster MP, Mason RT. 2001. Movements, mating,
and dispersal of red-sided gartersnakes (Thamnophis sirtalis parietalis) from a
communal den in Manitoba. Copeia 2001: 82-91.
Shine R, Langkilde T, Mason R. 2004. Courtship tactics in garter snakes: How do a male's
morphology and behaviour influence his mating success? Animal Behaviour 67: 477-483.
Shine R, Langkilde T, Mason RT. 2003. Cryptic forcible insemination: Male snakes exploit female
physiology, anatomy and behavior to obtain coercive matings. The American Naturalist
162: 653-667.
Shine R, Langkilde T, Wall M, Mason R. 2006a. Temporal dynamics of emergence and dispersal
of garter snakes from a communal den in Manitoba. Wildlife Research 33: 103-111.
Shine R, Langkilde T, Wall M, Mason RT. 2006b. Temporal dynamics of emergence and dispersal
of garter snakes from a communal den in Manitoba. Wildlife Research 33: 103-111.
Shine R, O'Connor D, Mason RT. 2000. Sexual conflict in the snake den. Behavioral Ecology and
Sociobiology 48: 392-401.
Shine R, Olsson MM, Mason RT. 2000. Chastity belts in gartersnakes: The functional significance
of mating plugs. Biological Journal of the Linnean Society 70: 377-390.
Shine R, Olsson MM, Moore I, LeMaster MP, Greene M, Mason RT. 2000. Body size enhances
mating success in male gartersnakes. Animal Behaviour 59: 4-F11.
Shine R, Phillips B, Langkilde T, Lutterschmidt DI, Waye H, Mason RT. 2004. Mechanisms and
consequences of sexual conflict in garter snakes (Thamnophis sirtalis, Colubridae)
Behav. Ecol. 654-660.
Shine R, Phillips B, Waye H, LeMaster M, Mason R. 2003. Chemosensory cues allow courting
male garter snakes to assess body length and body condition of potential mates.
Behavioral Ecology and Sociobiology 54: 162-166.
Shine R, Wall M, Langkilde T, Mason RT. 2005. Battle of the sexes: forcibly inseminating male
garter snakes target courtship to more vulnerable females. Animal Behaviour 70: 11331140.
Shine R, Webb JK, Lane A, Mason RT. 2006. Flexible mate choice: a male snake's preference for
larger females is modified by the sizes of females encountered. Animal Behaviour 71:
203-209.
Shuster SM, Wade MJ. 2003. Mating Systems and Strategies. Princeton University Press:
Princeton, New Jersey.
Siegel DS, Sever DM. 2006. Utero-muscular twisting and sperm storage in viperids.
Herpetological Conservation and Biology 1: 87-92.
Simmons L. 1987. Sperm competition as a mechanism of female choice in the field cricket,
Gryllus bimaculatus. Behavioral Ecology and Sociobiology 21: 197-202.
158
Simmons L, Parker G. 1992. Individual variation in sperm competition success of yellow dung
flies, Scatophaga stercoraria. Evolution 46: 366-375.
Simmons LW. 2001. Sperm Competition and Its Evolutionary Consequences in the Insects.
Princeton University Press: Princeton, New Jersey.
Simmons LW. 2005. The evolution of polyandry: Sperm competition, sperm selection, and
offspring viability. Annual Review of Ecology, Evolution, and Systematics 36: 125-146.
Simmons LW, Denholm A, Jackson C, Levy E, Madon E. 2007. Male crickets adjust ejaculate
quality with both risk and intensity of sperm competition. Biology Letters 3: 520-522.
Smith C, Pateman-Jones C, Zięba G, Przybylski M, Reichard M. 2009. Sperm depletion as a
consequence of increased sperm competition risk in the European bitterling, Rhodeus
amarus. Animal Behaviour 77: 1227-1233.
Smith C, Ryan M. 2010. Evolution of sperm quality but not quantity in the internally fertilized
fish Xiphophorus nigrensis. Journal of Evolutionary Biology 23: 1759-1771.
Smith RL. 1984. Sperm competition and the evolution of animal mating systems. Academic
Press: Orlando.
Snook RR, Chapman T, Moore PJ, Wedell N, Crudgington HS. 2009. Interactions between the
sexes: new perspectives on sexual selection and reproductive isolation. Evolutionary
Ecology 23: 71-91.
Sobel JM, Chen GF, Watt LR, Schemske DW. 2009. The biology of speciation. Evolution 64: 295315.
Stewart GR. 1972. An unusual record of sperm storage in a female garter snake (genus
Thamnophis). Herpetologica 28: 346-347.
Stockley P. 1997. Sexual conflict resulting from adaptations to sperm competition. Trends in
Ecology & Evolution 12: 154-159.
Stockley P. 1999. Sperm selection and genetic incompatibility: does relatedness of mates affect
male success in sperm competition? Proceedings: Biological Sciences: 1663-1669.
Suarez S, Pacey A. 2006. Sperm transport in the female reproductive tract. Human Reproduction
Update 12: 23.
Swanson WJ, Vacquier VD. 2002. The rapid evolution of reproductive proteins. Nat Rev Genet 3:
137-144.
Takami Y. 2003. Experimental analysis of the effect of genital morphology on insemination
success in the ground beetle Carabus insulicola (Coleoptera Carabidae). Ethology
Ecology & Evolution 15: 51-61.
Takami Y, Sasabe M, Nagata N, Sota T. 2008. Dual function of seminal substances for mate
guarding in a ground beetle. Behavioral Ecology 19: 1173-1178.
Thornhill R. 1983. Cryptic female choice and its implications in the scorpionfly Harpobittacus
nigriceps. American Naturalist: 765-788.
Thornhill R, Alcock J. 1983. The Evolution of Insect Mating Systems. Harvard University Press.
Tokarz RR. 1999. Relationship between copulation duration and sperm transfer in the lizard
Anolis sagrei. Herpetologica 55: 234-241.
Tourmente M, Cardozo G, Bertona M, Guidobaldi A, Giojalas L, Chiaraviglio M. 2006. The
ultrastructure of the spermatozoa of Boa constrictor occidentalis, with considerations on
its mating system and sperm competition theories. Acta Zoologica 87: 25-32.
159
Tourmente M, Cardozo G, Guidobaldi H, Giojalas L, Bertona M, Chiaraviglio M. 2007. Sperm
motility parameters to evaluate the seminal quality of Boa constrictor occidentalis, a
threatened snake species. Research in Veterinary Science 82: 93-98.
Tourmente M, Giojalas LC, Chiaraviglio M. 2011. Sperm parameters associated with
reproductive ecology in two snake species. Herpetologica 67: 58-70.
Tourmente M, Gomendio M, Roldan ERS, Giojalas LC, Chiaraviglio M. 2009. Sperm competition
and reproductive mode influence sperm dimensions and structure among snakes.
Evolution 63: 2513-2524.
Tregenza T, Wedell N. 2000. Genetic compatibility, mate choice and patterns of parentage:
Invited review. Molecular Ecology 9: 1013-1027.
Uller T, Olsson M. 2008. Multiple paternity in reptiles: patterns and processes. Molecular
Ecology 17: 2566-2580.
Uller T, Stuart-Fox D, Olsson M. 2010. Evolution of primary sexual characters in reptiles
Evolution of Primary Sexual Characters in Animals. Oxford: Oxford University Press. 425452.
Volsøe H. 1944. Structure and seasonal variation of the male reproductive organs of Vipera
berus B. Lunos Bogtrykkeri: Copenhagen.
Voss RS. 1979. Male accessory glands and the evolution of copulatory plugs in rodents. Museum
of Zoology, University of Michigan: Ann Arbor, Michigan.
Wade M, Patterson H, Chang N, Johnson N. 1994. Postcopulatory, prezygotic isolation in flour
beetles. Heredity 72: 163.
Wedell N, Cook PA. 1999. Butterflies tailor their ejaculate in response to sperm competition risk
and intensity. Proceedings of the Royal Society of London. Series B: Biological Sciences
266: 1033-1039.
Wedell N, Gage MJG, Parker GA. 2002. Sperm competition, male prudence and sperm-limited
females. Trends in Ecology & Evolution 17: 313-320.
Whittier JM, Crews D. 1986. Ovarian development in red-sided garter snakes, Thamnophis
sirtalis parietalis: relationship to mating. Gen Comp Endocrinol 61: 5-12.
Whittier JM, Mason RT, Crews D. 1985. Mating in the red-sided garter snake, Thamnophis
sirtalis parietalis: differential effects on male and female sexual behavior. Behavioral
Ecology and Sociobiology 16: 257-261.
Wigby S, Chapman T. 2004. Sperm competition. Current Biology 14: 100-103.
Wolfner MF. 2009. Battle and ballet: molecular interactions between the sexes in Drosophila.
Journal of Heredity 100: 399-410.
Wong A. 2011. The molecular evolution of animal reproductive tract proteins: what have we
learned from mating-system comparisons? International journal of evolutionary biology
2011.
Wusterbath T, King RB, Duvall MR, Grayburn WS, Burghardt GM. 2010. Phylogenetically
widespread multiple paternity in New World natricine snakes. Herpetological
Conservation and Biology 2010: 86-93.
Yasui Y. 1997. A" good-sperm" model can explain the evolution of costly multiple mating by
females. The American Naturalist 149: 573-584.
Yasui Y. 1998. Thegenetic benefits' of female multiple mating reconsidered. Trends in Ecology &
Evolution 13: 246-250.
160
Yasui Y. 2001. Female multiple mating as a genetic bet‐hedging strategy when mate choice
criteria are unreliable. Ecological Research 16: 605-616.
Zeh JA, Zeh DW. 1994. Last-Male Sperm Precedence Breaks down when Females Mate with
Three Males. Proceedings of the Royal Society B: Biological Sciences 257: 287-292.
Zeh JA, Zeh DW. 2001. Reproductive mode and the genetic benefits of polyandry. Animal
Behaviour 61: 1051-1063.